Endocrine Pathology (2019) 30:96–105 https://doi.org/10.1007/s12022-019-9578-3 Parathyroid Neoplasms: Immunohistochemical Characterization and Long Noncoding RNA (lncRNA) Expression Qiqi Yu1 & Heather Hardin1 & Ying-Hsia Chu1 & William Rehrauer1 & Ricardo V. Lloyd1 Published online: 22 May 2019 # Springer Science+Business Media, LLC, part of Springer Nature 2019 Abstract Parathyroid adenomas are slow growing benign neoplasms associated with hypercalcemia, while atypical parathyroid adenomas and parathyroid carcinomas are uncommon tumors and their histologic features may overlap with parathyroid adenomas. LncRNAs participate in transcription and in epigenetic or post-transcriptional regulation of gene expression, and probably contribute to carcinogenesis. We analyzed a group of normal, hyperplastic, and neoplastic parathyroid lesions to determine the best immunohistochemical markers to characterize these lesions and to determine the role of selected lncRNAs in tumor progression. A tissue microarray consisting of 111 cases of normal parathyroid (n = 14), primary hyperplasia (n = 15), secondary hyperplasia (n = 10), tertiary hyperplasia (n =11),adenomas(n = 50), atypical adenomas (n =7),andcarcinomas(n =4)was used. Immunohistochemical staining with antibodies against chromogranin A, synaptophysin, parathyroid hormone, and insulinoma-associated protein 1(INSM1) was used. Expression of lncRNAs including metastasis-associated lung adenocarcino- ma transcript one (MALAT1), HOX transcript antisense intergenic RNA (HOTAIR), and long intergenic non-protein coding regulator of reprograming (Linc-ROR or ROR) was also analyzed by in situ hybridization and RT-PCR. All of the parathyroid tissues were positive for parathyroid hormone, while most cases were positive for chromogranin A (98%). Synaptophysin was expressed in only 12 cases (11%) and INMS1 was negative in all cases. ROR was significantly downregulated during progression from normal, hyperplastic, and adenomatous parathyroid to parathyroid carcinomas. These results show that parathyroid hor- mone and chromogranin A are useful markers for parathyroid neoplasms, while synaptophysin and INSM1 are not very sensitive broad-spectrum markers for these neoplasms. LincRNA ROR may function as a tumor suppressor during parathyroid tumor progression. Keywords Parathyroid . In situ hybridization . INMS1 . lncRNA . Parafibromin Introduction [6] have assisted in the distinction between parathyroid adenomas and carcinomas. However, significant chal- Parathyroid carcinomas are difficult lesions for a patholo- lenges in making these distinctions remain [3]. gist to diagnose and to distinguish from parathyroid ade- Broad spectrum neuroendocrine markers such as nomas in some cases [1, 2]. Lesions with some of the chromogranin and synaptophysin have been used to char- histological features of parathyroid adenomas and carci- acterize neuroendocrine tumors with secretory granules nomas have been designated atypical parathyroid adeno- from many tissues and organs in the body [1]. These neu- mas [1, 3]. In recent years, the application of immunohis- roendocrine biomarkers have proven to be very effective tochemical staining for Ki-67 [4] and other biomarkers as general markers for the diagnosis of neuroendocrine such as retinoblastoma gene product [4], parafibromin tumors. In recent years, another broad spectrum neuroen- [5], and fluorescent in situ hybridization (FISH) studies docrine marker, insulinoma-associated one (INSM1), is a new marker for neuroendocrine cells and tumors [7–9]. Because INSM1 is a transcription factor, its localization * Ricardo V. Lloyd in the nucleus of the cells provides some obvious advan- [email protected] tages, such as low background staining, in the immuno- histochemical characterization of neuroendocrine tumors 1 Department of Pathology and Laboratory Medicine, University of Wisconsin School of Medicine and Public Health, 600 Highland Ave, [7]. A large series of parathyroid tissues and tumors have Madison, WI 53792, USA not been examined for INSM1 expression to date. Endocr Pathol (2019) 30:96–105 97 Noncoding RNAs range from microRNAs which average immunohistochemistry staining was performed using anti- around 22 nucleotides in length, to small nucleolar RNAs and parafibromin from Santa Cruz Biotechnology and used at Piwi-interacting RNAs [10]. Long noncoding RNAs 1:8000. All immunolabeled markers were visualized with (lncRNAs) are defined as transcripts that are 200 nucleotides DAB. Positive and negative controls were prepared with each or larger and do not contain a protein coding sequence. Many analysis. Staining was scored on the TMA based on the inten- studies have shown that there are thousands of lncRNAs sity of staining (0 negative, 1+ weak, 2+ moderate, and 3+ which may have many functions and may also contribute to strong) and on the percentage of positive cells (negative 0%, carcinogenesis as oncogenes or tumor suppressor [11–13]. focal 0–25%, and diffuse > 25%). Although microRNAs have been examined in parathyroid tis- sues by several investigators, very few studies of lncRNAs in In Situ Hybridization parathyroid tissues have been reported [14]. A recent study by Zhang et al. [14] reported expression of lncRNAs PNT1 and TMAs were probed for ROR, MALAT1, and HOTAIR expres- GLIS2-AS1 in parathyroid neoplasms, and they suggested sion using the RNAscope 2.5 HD-Brown Manual Assay that these molecules could be involved in parathyroid carci- (Advanced Cell Diagnostics, Newark, CA) as per manufac- noma tumorigenesis. We examined a series of normal, hyper- turer’s recommendations with the following modifications: an- plastic, and neoplastic parathyroid tissues using a tissue mi- tigen retrieval was performed in a Biocare Decloaker for 3 min, croarray (TMA) to characterize the immunohistochemical fea- protease digestion for 30 min, and probe incubation overnight tures and lncRNA expression patterns in benign and malig- at 40°C. The probes used included hs-ROR (315401), hs- nant parathyroid tissues. MALAT1 (400811), hs-HOTAIR (312341), hs-PPIB (positive control, 313901), and dapB (negative control, 310043) (Advanced Cell Diagnostics). Expression levels of all probes Materials and Methods were visualized with DAB as previously reported [15]. Tissue Microarrays Automated Multispectral Image Quantitation A tissue microarray (TMA) was constructed and included The hybridized TMA slides were visualized with the Vectra normal parathyroid tissues (N = 14) primary hyperplasia slide scanner (PerkinElmer, Waltham, MA). InForm 1.4.0 (n = 15), secondary hyperplasia (n = 10), tertiary hyperplasia software (PerkinElmer, Waltham, MA) was used to segment (n = 11), adenomas (n =50),atypicaladenomas(n = 7), and each tissue core into architectural compartments (tumor cells carcinomas (n = 4). One carcinoma consisted of three local versus stroma) and subcellular compartments (nucleus versus recurrences and a lung metastasis. Cases were identified by cytoplasm). Expression levels of ROR, MALAT1, and searching the institutional electronic medical record system HOTAIR were quantitated for each tissue core as average and confirmed by immunohistochemical analysis (details pro- nuclear optical density (OD) as previously described [15]. vided below). Each case was represented by duplicate 0.6-mm Results from duplicate cores were averaged for each case to cores which were prepared using a manual tissue arrayer obtain a representative score. (Beecher Instruments, Sun Prairie, WI). The study was ap- proved by the Institutional Review Board at the University Real-time PCR of Wisconsin-Madison. Total RNA was extracted from samples with TRIzol reagent Immunohistochemistry (ThermoFisher Scientific, Waltham, MA) according to the manufacturer’s instructions, and RNA quality and concentra- Immunohistochemical analyses were performed on a Ventana tions were assessed with a NanoDrop 1000 spectrophotometer BenchMark Ultra system (Ventana Medical Systems, Inc., (ThermoFisher Scientific, Waltham, MA). One microgram of Tucson, AZ) according to the manufacturer’s protocols. total RNA was reverse-transcribed using the All-in-One Primary antibodies used included Ki-67 (1:50 dilution with miRNA RT-qPCR detection kit (GeneCopoeia, Rockville, Van Gogh Yellow; Biocare, Pacheco, CA), parathyroid hor- MD). RT-qPCR was performed on a CFX96 PCR detection mone Clone MRQ-31, prediluted, Cell Marque, Rocklin,CA system (Bio-Rad Laboratories, Hercules, CA) using Bullseye chromogranin A (clone LK2H10, prediluted; Ventana EvaGreen qPCR master mix (MIDSCI, St. Louis, MO), nor- Medical Systems, Inc., Tucson, AZ), synaptophysin (poly- malized to 18S rRNA; relative fold change was determined by clonal,prediluted;CellMarque,Rocklin,CA),and the ΔΔ CT method as previously reported [15]. The PCR insulinoma-associated protein one (INSM1; clone SC- primers used for, ROR, HOTAIR, MALAT1, and 18S are as 271408 from Santa Cruz Biotechnology, Dallas, TX; 1:1000 follows: ROR; Forward 5′-CTGGCTTTCTGGTTTGACG-3′ dilution from Biocare, Pacheco, CA). In addition, manual and Reverse 5′-CAGGAGGTTACTGGACTTGGAG-3′, 98 Endocr Pathol (2019) 30:96–105 Fig. 1 Representative H&E sections of normal parathyroid (a), parathyroid adenoma (b), atypical parathyroid adenoma (c), and parathyroid carcinoma (d). The atypical adenomas and carcinomas had larger nuclei than the normal parathyroid and adenomas HOTAIR; Forward 5′-CAGTGGGGAACTCTGACTCG-3′ Statistical Analysis and Reverse 5′-GTGCCTGGTGCTCTCTTACC-3′, MALAT1; Forward 5′-GACGGAGGTTGAGATGAAGC-3′ Student’s t test