1. Parathyroid Cancer: Diagnosis and Utility of Frequent Biomarkers
As previously mentioned, the definitive diagnosis of PC requires unequivocal criteria for invasion in the postoperative biopsy (angioinvasion, lymphoinvasion, perineural invasion, invasion of neighboring soft tissues and musculoskeletal structures, thyroid, esophagus, etc. and/or presence of lymph node or distant metastasis). The mentioned criteria are known as absolute criteria, and meeting just one is sufficient for the diagnosis of PC. On the other hand, the cytological or histological characteristics of the tumor are not sufficient to make the diagnosis, which determines the difficulty in discriminating between PT adenoma and carcinoma. Criteria associated with malignancy can be reviewed in the Introduction section [
3,
16]. Given the difficulty in confirming the diagnosis of PC, attempts have been made to address this problem by searching for biomarkers mainly through IHC; however, none have been found to be pathognomonic of the disease [
16].
In conclusion, due to the low frequency of this cancer, most biomarker studies in PC are in early phases, do not have clear definitions regarding results, show variations in populations used and reporting methods, and also have a small sample size (1-57 cases per study), which has made it difficult to perform statistical analyses. In this context, a recent systematic review of potential biomarkers in PC by Davies et al. significantly identified (in at least one study) 5 markers in serum and 13 in tissue. Specifically, the markers that showed the strongest evidence in serum are calcium and PTH (15 and 16 studies, respectively), with a cutoff of 3 mmol/L (12.02 mg/dL) and >3 times the upper limit of normal (ULN) for calcium and PTH, respectively, in order to suspect PC. In addition, the presence of elevated alkaline phosphatase or creatinine, as well as decreased 25-hydroxyvitamin D, are independently associated with suspicion of PC. It is likely that these latter markers are associated with a biochemical manifestation of disease severity rather than directly related to the pathophysiology of the disease (as is the case with serum calcium and PTH) [
20].
On the other hand, multiple biomarkers have been described in tissue, which implies access to tissue and post-surgical analysis by the pathologist. Despite the existence of several markers, none of them are pathognomonic of PC and it is currently an active area of research, where every study is an advancement.
Parafibromin is the main biomarker studied in tissues; it is a tumor suppressor protein, encoded by the Cell Division Cycle 73 (CDC73, previously HRPT2) gene. Its function lies in the repression of cyclin D1, consequently stopping the cell cycle, and it is also associated with the regulation of the P53 pathway. Mutations in this gene are associated with decreased function and detection by IHC [
16]. As a result, the absence of parafibromin is associated with diagnostic evidence and worse prognosis in PC. Over 19 scientific articles have described its association with PC, showing significant results in 7 of them [
20]. Despite the above, there is an overlap in the presence of this gene and protein between PC and adenomas, as CDC73 mutations have been described in 1-6% of adenomas. Given the low frequency of this cancer, a nearly perfect specificity is required for its use as a screening tool in PT tumors (as a standalone marker), which complicates its use in this aspect. However, it is a strong tool for ruling out malignancy in case of positive immunoreactivity. It should be used while considering the clinical context of the patient, and currently it is associated with other markers [
21].
Galectin 3 is a glycoprotein involved in pathophysiological processes such as fibrosis, inflammation, and cancer. It plays a regulatory role in the tumor microenvironment, specifically by suppressing T cells through inhibition of T-cell receptor (TCR)-mediated signaling. Its overexpression has been associated with tumor growth and invasion [
22]. Six scientific articles have associated the overexpression of Galectin 3 with PC, but there are reports that raise doubts about the utility of this marker [
20]. Recently, Mohammed et al. reported only 15% positivity in PC, showing a poor profile as a biomarker: sensitivity (Sen) 6%, specificity (Spe) 29%, positive predictive value (PPV) 19%, and negative predictive value (NPV) 10% [
23]. Further studies are needed for this biomarker.
Ki67 is a nuclear protein expressed in proliferating cells (active phase of the cell cycle) and downregulated during the G0 phase. Therefore, increased staining for this antigen suggests a higher proportion of proliferative cells in the histological sample. It is a diagnostic and prognostic tool used in multiple types of carcinomas (especially colorectal), where staining >5% may aid in the diagnosis of PC, although it does not yet qualify as a diagnostic criterion (thus, it is used in association with other markers). However, there are reports that do not show significant differences between adenomas and PC, indicating the need for further studies [
16,
20,
24].
PGP9.5 (protein gene product 9.5) is the protein product of UCHL1 (ubiquitin carboxyl-terminal esterase L1) and is mainly expressed in nervous and neuroendocrine tissues. Its presence has been associated with aggressive neoplasms (colorectal, prostate, gastric, and pulmonary). There are at least two studies where PGP9.5 is reported to be overexpressed in PC, but further studies are needed to validate these findings [
20,
25].
APC (
Adenomatous polyposis coli) is a tumor suppressor gene that acts by inhibiting the Wnt signaling pathway. It can be evaluated through polymerase chain reaction (PCR) or its protein product via IHC, and its loss of expression has been associated with carcinogenesis. Currently, it is also linked to other carcinomas such as colon and liver cancer. At least two studies have shown a significant association with PC, but, similar to other markers, there are articles that have not demonstrated this association, emphasizing the need for further studies on this topic [
20].
The protein p27 acts as a tumor suppressor and a cyclin-dependent kinase inhibitor, which slows down the cell cycle. Loss of p27 expression is suggestive of PC, with at least one statistically relevant study supporting this association [
16,
20].
The decreased expression of the calcium-sensing receptor (CaSR) has been reported in PC, which is a rare phenomenon in benign PT tumors. However, the reports on this association are inconsistent, which makes it challenging to use CaSR as a reliable biomarker. Filamin A is a scaffold protein that binds to the CaSR and facilitates the activation of the MAPK pathway. In a study by Storvall et al., cytoplasmic expression of Filamin A was significantly higher in PC compared to adenomas, suggesting its potential as a biomarker for PC [
16,
26].
Table 1.
Main markers studied in PC.
Table 1.
Main markers studied in PC.
Marker |
Source |
Characteristics |
References |
Calcium |
Serum |
Increased |
[27,28] |
PTH |
Serum |
Increased |
[27,28,29,30] |
Alkaline phosthatase |
Serum |
Increased |
[27,28,31,32] |
Vitamin D |
Serum |
Decreased |
[33] |
Parafibromin |
Tissue |
Mutated/decreased |
[23,33,34,35,36,37] |
Galectin 3 |
Tissue |
Increased |
[23,37,38,39,40] |
Ki67 |
Tissue |
Increased |
[23,30,33,37,40] |
PGP 9.5 |
Tissue |
Increased |
[37,41] |
APC |
Tissue |
Decreased |
[23,42] |
P27 |
Tissue |
Decreased |
[43,44] |
CaSR |
Tissue |
Decreased |
[16,26] |
Apart from the markers mentioned above, there are limited studies on other markers such as: AgNOR, CDC73, cyclin D1, P21, P53, Rb (retinoblastoma gene), BCL2, BAX, Mdm2, PTEN, among others. Most of these are individual studies and often lack statistical analysis. Undoubtedly, this is an area of research that is still not well developed but holds great promise for the future. It is important to note that no single biomarker has been able to differentiate PC from adenomas on its own; but rather these markers can serve as auxiliary tools in diagnosis and prognosis [
3,
20].
2. Epithelial-Mesenchymal Transition and Cancer Stem Cells in Parathyroid Carcinoma
The presence of CSC and patterns of dedifferentiation, such as EMT, have been established in various types of cancer. Specifically, EMT is a cellular process in which an epithelial cell acquires a mesenchymal phenotype and behavior due to the downregulation of specific epithelial markers (such as cytokeratin and E-cadherin) and upregulation of mesenchymal markers (such as fibronectin, N-cadherin, and vimentin). Consequently, these cells acquire a fibroblast-like morphology, lose apico-basal polarity, as well as, interaction with the basement membrane and therefore exhibit increased migratory capacity and invasive properties [
45]. This process is regulated by complex molecular pathways, and can be described based on certain molecular markers. The transcription factors (EMT-TFs) involved in this process include Zeb1, Zeb2, Snail1, Snail2 (Slug), and Twist1, which are responsible for silencing genes expressed in epithelial cells and inducing those specific to mesenchymal cells [
46].
At the post-translational level, increased expression of certain proteins such as N-cadherin, vimentin, fibronectin, matrix metalloproteinases (-2, -3, -9) and downregulation of E-cadherin, cytokeratin, occludins, desmoplakin, among others, are canonical events in this cellular process [
47].
On the other hand, Cancer Stem Cells (CSCs) or tumor-initiating cells are cancer cells that share characteristics with stem or progenitor cells. Of particular note is their ability to differentiate into other cell lineages and undergo self-renewal through symmetric (generating two stem cells) or asymmetric (generating one stem cell and one differentiated cell) cell division. CSCs are involved in tumour formation, invasion, metastasis and even resistance to chemotherapy and radiotherapy.. There are multiple methods for detecting CSCs, with the characterization of CSCs using cellular biomarkers being a prominent approach. These biomarkers typically correspond to membrane antigens and transcription factors, and may vary depending on whether the tumors are solid or hematological cancers. In solid tumors, prominent surface markers include CD24, CD44, CD90, CD133, and EpCAM, while intracellular markers such as Sox2, Oct-3/4, and Nanog are also used, among others [
48,
49].
There are limited reports specifically aimed at evaluating these phenomena in PC. According to a recent review by Uljanovs et al., the low incidence of PC has hindered studies on epithelial-mesenchymal transition in this cancer [
16].
Fendrich et al. compared the expression of EMT markers (E-cadherin, Snail, and Twist) using IHC in normal PT tissue (2 cases) and PT adenomas (25 cases), hyperplasia (25 cases), and PC (9 cases). In all cases of PC, loss of membranous staining of E-cadherin was observed, showing a characteristic cytoplasmic staining pattern indicative of EMT; while membranous staining was preserved in normal tissues and benign proliferative disorders. On the other hand, Snail and Twist lost the homogeneous expression pattern seen in the control groups, being limited to the invasive front in the cancerous tissue samples. These changes in marker expression observed in PC suggest an important role of EMT in the tumorigenesis of this cancer. Although further studies are needed, this could be a potential tool to distinguish between PT adenomas and carcinomas [
50].
Schneider et al. further supported the previously described change in immunostaining pattern and compared the expression of E-cadherin in atypical PT adenomas (currently known as atypical parathyroid tumors), benign PT adenomas, and PC. It was concluded that atypical tumors showed membranous staining of E-cadherin, characteristic of benign proliferative disorders of PT tissue, in contrast to PC. The potential use of E-cadherin as a differentiating marker between these two clinical entities was suggested [
51].
Meanwhile, Pandya et al. conducted a genomic profiling of 17 cases of PC using whole-exome sequencing. Among the results, recurrent somatic mutations were described in the Zeb1 gene (3 cases), a transcriptional repressor associated with tumor invasion and metastasis through the activation of EMT [
52].
In addition to the previously mentioned biomarkers, the study of intermediate filaments has shown that cytokeratin 19 presents a progressive upregulation in proliferative PT lesions (adenoma, hyperplasia, and cancer), statistically significant but markedly heterogeneous. Further studies of this intermediate filament in PC are recommended. On the other hand, vimentin corresponds to a mesenchymal intermediate filament with important functions in cell mobility, signaling, and migration; however, it is difficult to study in PT tissue due to its glandular histology. Recently, Ulijanovs et al. described that the fraction of parenchymal cells positive for vimentin varies in healthy tissue, hyperplasia, adenoma, and carcinoma, with percentages of 9%, 11.7%, 19.3%, and 36.8% respectively. Regarding location, in healthy tissue, vimentin expression is only perinuclear, while carcinoma shows significant cytoplasmic expression, and adenomas and hyperplasias show a combination of both patterns [
16,
53].
On the other hand, CD44, a surface glycoprotein and a recognized CSC marker, has been less studied in PT pathology, and currently, is not believed to play a significant role in PT tumors given its absence in normal PT tissue and proliferative disorders. This is interesting considering that, generally, neuroendocrine tumors express this marker. This discrepancy could possibly be explained by the fact that PT tissue has an embryological origin in the neural crest, which is CD44(-), whereas neuroendocrine tumors that are CD44(+) originate from the endoderm. This could also explain the presence of E-cadherin, Snail, and Twist in normal and benign proliferative PT tissue. Based on this, it would be interesting to evaluate markers derived from the neural crest that may be useful in PC. Additionally, CD56, a membrane glycoprotein and a member of the immunoglobulin superfamily, has been described as a negative marker in PC, despite being commonly seen in neuroendocrine tumors of other organs. CD56 is not expressed in normal PT tissue or proliferative disorders, and its presence can help rule out a PT origin of the tumor [
16,
53].
3. Parathyroid Carcinoma Genomics
Advances in genomics are emerging and have enormous potential as they could be used relatively simply and reliably as diagnostic tools or decision aids.
Firstly, the CDC73 gene (also known as HRPT2) is a tumor suppressor gene that encodes a protein called parafibromin (whose function has been previously mentioned). Multiple somatic and germline mutations associated with partial or total inactivation of this gene have been identified. Somatic mutations in this gene are estimated to be the main genetic alteration in sporadic PC, present in two-thirds of cases (reports vary between 9-70%). Germline mutations are associated with HPT-JT syndrome, a population with a 20% probability of developing PC [
54,
55].
Secondly, the PRUNE2 gene encodes the homonymous protein, which has tumor suppressor activity by suppressing the activity of RhoA, inhibiting oncogenic transformation. Recently, Yu et al. detected several mutations via whole-exome sequencing in 18% of patients with PC, but did not describe this mutation in PT adenomas. This could represent PRUNE2 as a marker that differentiates between both conditions [
55,
56].
Furthermore, the CCND1 gene acts as an oncogene in PT adenomas through the PTH-CCND1 rearrangement. This gene encodes cyclin D1, whose amplification has been observed in up to 71-90% of PC and is associated with cell proliferation. Recent studies have shown that amplification of CCND1 can be mutually exclusive with somatic mutations in CDC73, suggesting alternative mechanisms of malignancy [
52,
55].
On the other hand, multiple endocrine neoplasia type 1 (MEN1) is an autosomal dominant endocrine disorder caused by mutations in the MEN1 gene. While this syndrome is commonly associated with PT adenomas or hyperplasia, around 1% of cases may present with PC. Recent studies indicate that mutations in this gene may be present in approximately 13 to 31% of cases of PC, so additional studies could position it as a marker of the disease [
55].
Genes associated with the PI3K/AKT/mTOR and Wnt signaling pathways have also been implicated in PC. The first signaling pathway plays a physiological role in cell growth and survival, and its overactivation has been associated with various types of cancers, particularly through increased cell growth, angiogenesis, suppression of senescence and autophagy, as well as increased metastatic potential. Mutations in the PI3K/AKT/mTOR pathway have been described in 13-20% of PC cases, making it a potential target for inhibitors of this pathway [
55,
57].
Similarly, the Wnt signaling pathway is physiologically associated with cell proliferation, survival, and apoptosis. Aberrant activation of this pathway has also been implicated in multiple neoplasms. It has been described that CDC73 regulates this pathway through stabilization of beta-catenin. In addition, recent studies have identified a regulatory role of this pathway through genetic and epigenetic changes in APC and RNF43, observed in rare cases of PC as well as colorectal and endometrial carcinomas [
55].
Finally, somatic mutations have been described in other genes such as TERT, which is associated with telomerase activity through the coding of its catalytic subunit. Additionally, mutations have been identified in genes AKAP9 (signal transduction pathway), FAT3 (cellular adhesion), and ZEB1 (EMT), the latter of which was mentioned previously [
55].
4. Epigenetics: miRNAs, lncRNAs, circRNAs
Epigenetics plays a crucial role in tumorigenesis. It encompasses all intracellular mechanisms involved in the modulation of gene expression without altering the DNA sequence. Within this field, the action of non-coding RNAs (ncRNAs) at the post-transcriptional level is noteworthy. These ncRNAs can be classified based on length into small RNAs (sRNAs) or long non-coding RNAs (lncRNAs), with lengths less than or greater than 200 nucleotides, respectively. Specifically, the former group includes microRNAs (miRNAs), while the latter encompasses a circular group of lncRNAs called circRNAs. Similar to tissue biomarkers, ncRNAs have been used to differentiate adenomas from carcinomas, with variable results, and so far, no pathognomonic marker for the disease has been identified [
58].
MiRNAs (or miRs) regulate gene expression at the post-transcriptional level, usually through negative regulation of mRNA, specifically by binding to complementary sites (typically in the UTR region). These miRNAs have different expression patterns in tissues and serum, with the latter location providing the theoretical basis for liquid biopsies, with potential applications in non-invasive diagnosis and monitoring. Specifically, in the context of the PT gland, miRNAs participate in the regulation of hormonal synthesis and secretion, as well as tumorigenesis [
59].
At the serum level, Wang et al. described a significant up-regulation of MiR-27a-5p in serum exosomes of patients with PC; revealing an area under the curve (AUC) of 0.859. It is relevant that this specific miRNA participates in the activation of the Wnt/β-catenin signaling pathway, playing a key role in the process of EMT [
60].
On the other hand, Krupinova et al. reported that miR-342-3p is significantly decreased in PC, with an AUC of 0.89 when used as a biomarker [
59,
61].
Studies evaluating miRNA expression at the tissue level have shown variable results. Corbetta et al. found that miR-296 (downregulated in PC), miR-222, and miR-503 (upregulated in PC) varied significantly between adenomas and PC. On the other hand, higher levels of HGS mRNA (hepatocyte growth factor receptor-regulated tyrosine kinase substrate) were observed in PC, which has been associated with EMT-related phenomena such as downregulation of E-Cadherin and subsequent increase in invasion and metastasis. Interestingly, HGS mRNA is a direct target of miR-296, and miR-296 undergoes significant downregulation during tumor progression. Consequently, an indirect role of miR-296 in the carcinogenesis of PT is suggested [
62].
In parallel, Rahbari et al. reported significant variation of three miRNAs in PC: miR-26b, miR-30b, and miR-126. Notably, miR-126 showed better characteristics as a biomarker, with an AUC of 0.776 [
63].
On the other hand, Vaira et al. described that aberrations in the expression of miRNAs from the C19MC cluster are a characteristic of PC compared to PT adenomas. This cluster of miRNAs is usually silenced by hypermethylation in adults, and its association with various human tumors has been described, promoting tumor invasion and metastasis. Specifically within this cluster, miR-517c showed the most notable difference, correlating with tumor weight, serum calcium, and PTH levels [
64].
Additionally, increased expression of miRNAs from the C19MC cluster and miR-372 has been identified in PC distant metastases [
59]. Finally, Hu et al. report that in a Chinese population (which has a higher incidence of PC, accounting for 5-7% of HPTP cases), the following miRNAs showed differential expression between PC and PT adenomas: miR-139, miR-222, miR-30b, miR-517c, and miR-126. Specifically, the combination of miR-139 and miR-30b was found to be the best biomarker with an AUC of 0.89 [
59,
65]. It is important to highlight that most of these studies are limited by small sample sizes due to the low frequency of this cancer.
On the other hand, lncRNAs and circRNAs primarily act as endogenous competitive RNAs (binding to miRNAs and inhibiting their function), while circRNAs can also be translated into functional micropeptides. There are limited studies on these ncRNAs in PC. In brief, Jiang et al. described that a profile composed of the lncRNAs: LINC00959, lnc-FLT3-2:2, lnc-FEZF2-9:2, and lnc-RP11-1035H13.3.1-2:1 had an AUC of 0.88, with Sen: 81.8% and Spe: 83.9% for differentiating adenomas from PC [
66].
Furthermore, Zhang et al. recently described that alterations in the expression of lncRNAs PVT1 and GLIS2-AS1 had AUC values of 0.871 and 0.860, respectively [
67]. On the other hand, Hu et al. determined that circRNA_0075005 had an AUC of 0.77 for the same function [
58,
65].
7. Use of Panels
Given that none of the mentioned markers are pathognomonic of parathyroid cancer, the creation of panels, nomograms or other systems that allow for the combination of different markers has become common. These instruments offer an increased sensitivity, specificity, positive predictive value, and negative predictive value therefore maximizing the diagnostic performance of these markers.
Briefly, we present some examples: Kumari et al. describe that a panel combination of parafibromin, galectin-3, and PGP9.5 presents a Sen: 50%, Spe: 97.9%, and a predictive accuracy of 95.4% [
37] On the other hand, Truran et al. determined that the use of an immunohistochemical panel using the same markers described previously plus Ki67 is better than the isolated use of any single marker (Sen: 79% and Spe: 100%). These authors point out that a positivity greater than 5% of the cells to Ki57 is considered suspicious of being PC [
70]. Finally, Silva-Figueroa et al. developed a nomogram using IHC biomarkers: parafibromin, Rb, Ki67, E-cadherin, and galectin-3, which showed an AUC of 84.9% (adjusted AUC for optimism: 80.5%) [
71].