Profiling analysis of long non-coding RNA and mRNA in parathyroid carcinoma

in Endocrine-Related Cancer
Correspondence should be addressed to Q Liao or Y Zhao: or
Restricted access

Parathyroid carcinoma (PCa) is a rare endocrine neoplasia that typically has unfavourable outcomes. The contribution of long non-coding RNAs (lncRNAs) to the development of malignant and benign parathyroid tumours remains largely unknown. In this study, we explored transcriptomic profiling of lncRNA and mRNA expression in 6 PCa, 6 parathyroid adenoma (PAd) and 4 normal parathyroid (PaN) tissues. In total, 2641 lncRNA transcripts and 2165 mRNA transcripts were differentially expressed between PCa and PAd. Enrichment analysis demonstrated that dysregulated transcripts were involved mainly in the extracellular matrix (ECM)–receptor interaction and energy metabolism pathways. Bioinformatics analysis suggested that ATF3, ID1, FOXM1, EZH2 and MITF may be crucial to parathyroid carcinogenesis. Series test of cluster analysis segregated differentially expressed lncRNAs and mRNAs into several expression profile models, among which the ‘plateau’ profile representing components specific to parathyroid carcinogenesis was selected to build a co-expression network. Seven lncRNAs and three mRNAs were selected for quantitative RT-PCR validation in 16 PCa, 41 PAd and 4 PaN samples. Receiver-operator characteristic curves analysis showed that lncRNA PVT1 and GLIS2-AS1 yielded the area under the curve values of 0.871 and 0.860, respectively. Higher hybridization signals were observed in PCa for PVT1 and PAd for GLIS2-AS1. In conclusion, the current evidence indicates that PAd and PCa partially share common signalling molecules and pathways, but have independent transcriptional events. Differentially expressed lncRNAs and mRNAs have intricate interactions and are involved in parathyroid tumourigenesis. The lncRNA PVT1 and GLIS2-AS1 may be new potential markers for the diagnosis of PCa.

Downloadable materials

  • Supplementary Table 1 Clinical and laboratory characteristics of parathyroid carcinoma cohort
  • Supplementary Table 2 Diagnostic efficacy of PVT1 and GLIS2-AS1 in differentiating CDC73-mutant PCa via ROC analysis
  • Supplementary Figure 1 IPA molecular network 1
  • Supplementary Figure 2 IPA molecular network 2
  • Supplementary Figure 3 IPA molecular network 3
  • Supplementary Figure 4 The predicted regulatory effects via IPA analysis. Integrated analysis of up-stream regulators, biological functions and gene interactions revealed the most focused gene crosstalk. Orange ellipses and lines represent predicted activation and leading to activation. Blue ellipses and lines represent predicted inhibition and leading to inhibition. Yellow lines represent findings that are inconsistent with state of down-stream molecules. Grey lines indicate unpredicted effects.
  • Supplementary Figure 5 The predicted core up-stream regulators (ATF3)
  • Supplementary Figure 6 The predicted core up-stream regulators (ID1)
  • Supplementary Figure 7 The predicted core up-stream regulators (FOXM1)
  • Supplementary Figure 8 The predicted core up-stream regulators (EZH2)
  • Supplementary Figure 9 The predicted core up-stream regulators (MITF)


      Society for Endocrinology

Related Articles

Article Information


All Time Past Year Past 30 Days
Abstract Views 1144 1144 119
Full Text Views 86 86 12
PDF Downloads 40 40 2



  • View in gallery

    Differentially expressed lncRNAs and mRNAs in parathyroid carcinoma (PCa) and adenoma (PAd) tissues. (A) Venn diagrams present significantly changed profiles of PCa and PAd compared with PaN (FC ≥ 2, P < 0.05), showing overlaps between PCa- and PAd-specific genes. (B and C) Volcano plots of mRNA (B) and lncRNA (C) expression variations between PCa and PAd. (D and E) Heat map of 2165 differentially expressed mRNAs (D) and 2641 lncRNAs (E) between PCa and PAd (FC ≥ 2, P < 0.05). Red indicates up-regulated transcripts, while green indicates downregulated transcripts. (F) Categories and numbers of 2641 differentially expressed lncRNAs. PAd, parathyroid adenoma; PaN, normal parathyroid glands; PCa, parathyroid carcinoma. A full colour version of this figure is available at

  • View in gallery

    The results of GO and KEGG pathway analysis. (A and B) The horizontal axis represents the enrichment factor, and the vertical axis represents the GO or pathway category. (C) Leading pathway annotation indicates the ECM-receptor interaction pathway. Red marks are associated with upregulated genes, while the nodes in brilliant green indicate down-regulation. Light green indicates no significance. GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

  • View in gallery

    IPA analysis of differentially expressed mRNA between PCa and PAd. One of the most significant networks with 114 focus molecules (score, 96). Upregulated node genes are depicted in red, while downregulated genes are depicted in green. A full colour version of this figure is available at

  • View in gallery

    STC analysis of expression profiles of lncRNAs and mRNAs in parathyroid tumours. (A) Each box represents a model expression pattern. The upper number ranging from 0 to 15 marks the mode profile. A total of six significant profiles were identified with P < 0.05. (B) Profiles 7 and 8 are the most significant patterns, which were defined as ‘plateau’ type. (C) Profiles 2 and 13 were defined as ‘coherent’ type. (D) Expression patterns of profiles 6 and 9 are presented. The vertical axis shows the transcript expression level after Log2 normalized transformation. A_N, normal tissue; B, parathyroid carcinoma; D, parathyroid adenoma. A full colour version of this figure is available at

  • View in gallery

    Co-expression network of ‘plateau’ profile. In the network, continuous lines indicated gene correlations, ellipse nodes represented mRNAs and rectangle nodes represented lncRNAs. Genes in different profiles are shown as different colours (profile 7, violet; profile 8, pink). A degree scoring was used to evaluate hub genes in the network. Values of degree were calculated according to the connection among other genes. Larger size of the nodes indicated higher degree scoring and a more central role of lncRNA or mRNA. A full colour version of this figure is available at

  • View in gallery

    Validation of differentially expressed mRNAs and lncRNAs in PCa and PAd patients. (A, B and C) Selected mRNAs were validated by RT-qPCR. Upregulated (D, E, F, G and H) and downregulated (I and J) candidate lncRNAs were chosen for potential markers. All transcripts were verified in a cohort of PCa (n = 16), PAd (n = 41) and PaN tissue samples (n = 4). *P < 0.05, **P < 0.01. A full colour version of this figure is available at

  • View in gallery

    Diagnostic value of candidate lncRNAs assessed by ROC curve. Upregulated (A) and downregulated (B) lncRNAs that showed maximum AUC values were lncRNA PVT1 and lncRNA GLIS2-AS1. AUC, area under the curve.

  • View in gallery

    Representative ISH localization of lncRNA PVT1 and GLIS2-AS1 in parathyroid tumours (magnification: 200×). Representative images of lncRNA PVT1 and GLIS2-AS1 ISH signals in PCa and PAd are presented. The blue staining on FFPE tissues indicated positive ISH detection. A full colour version of this figure is available at

Cited By


Google Scholar