Protein Signatures of Parathyroid Carcinomas Using Proteomic Analyses

Abstract

Parathyroid carcinomas are rare endocrine tumors derived from the parathyroid glands with poor prognosis. Moreover, parathyroid carcinomas are resistant to radiation or drug therapy, and surgical resection is the only treatment option. Understanding the molecular pathogenesis of parathyroid carcinomas may pave the way for early diagnostic biomarkers and therapeutic targets. Therefore, we aimed to elucidate the protein signatures for parathyroid carcinomas through quantitative proteomic analyses. We performed liquid chromatography with tandem mass spectrometry (LC-MS/MS) technique with formalin-fixed paraffin-embedded (FFPE) samples and reached a quantitative depth of more than 5,000 proteins per sample. For the analyses, 23 parathyroid carcinoma and 15 adenoma samples were collected from five tertiary hospitals in Korea. Patients’ mean age was 52 years, and 24 (63%) were female. Patients with parathyroid carcinoma had higher parathyroid hormone (PTH) and serum calcium level than adenomas (PTH, 1077.6 ± 760.7, 181.8 ± 139.8 pg/mL; calcium 13.0 ± 2.8, 11.3 ± 1.0 mg/dL, respectively). From the proteomic expression profiling, there were 137 differentially expressed proteins with the cutoff of both p <0.05 and fold change > 1.5. Using the Ingenuity Pathway Analysis (IPA), top enriched canonical pathways in parathyroid carcinomas included glycoprotein-6 signaling related to the coagulation pathway, acute phase response, mTOR, and clathrin-/caveolar-mediated endocytosis signaling. In transcription factor analysis, TGFβ and TP53 were activated in carcinoma, and these factors were up-regulators of CD44 antigen and Annexin A2 (ANXA2) proteins. In network analysis, α-1-acid glycoprotein 1 (ORM1), laminin subunit β-2 (LAMB2), and Serpin family (SERPIN) proteins were derived as essential proteins and correlated to the AKT complex. Also, with the support vector machine (SVM)-based classification method, we derived a set of proteins that can discriminate carcinomas from adenomas, which consists of Carbonic anhydrase 4 (CA4), α/β hydrolase domain-containing protein 14B (ABHD14B), CD44, LAMB2, phosphatidylinositol transfer protein β isoform (PITPNB), and ORM1, with the lowest error rate of 11.1%. In conclusion, from the proteomics analyses of parathyroid neoplasms, newly recognized pathways - signaling related to coagulation, acute phase response, and endocytosis - were enriched in parathyroid carcinoma in addition to the known mTOR signaling pathway. The proteins such as α-1-acid glycoprotein and laminin subunit β-2 from SVM classification and network analyses could be the distinctive signature of carcinoma and may provide insights into the therapeutic target.