THCA is the most common form of endocrine system tumor, and its incidence has rapidly increased over the past few decades. PTC is the most predominant pathological subtype among all the THCA cases. Most PTC patients usually have a favorable prognosis when surgical treatment, radioactive iodine therapy, and thyroid-stimulating hormone (TSH) suppression treatment are implemented. Nonetheless, more than 10% of PTC patients may suffer local recurrence and distant metastasis after the initial treatment. Therefore, it is essential to explore new potential molecular targets for clinical therapy in order to improve patient outcomes.
TME consists of immune cells and non-immune stromal cells, and immune system was found to make a crucial contribution to cancer development and progression. Cancer immunotherapy has made tremendous progress for some cancer types in recent years.
Predictive or prognostic biomarkers related to the TIME have great promise for identifying novel molecular therapeutic targets and improving cancer patients clinical management of immunotherapy31.
Long non-coding RNAs (lncRNAs) have emerged as enormous amount and diverse functions in the past decade and have been reported to play important roles in a variety of biological processes, including cell proliferation, death, and tumor growth32. Enhancer was described as distal regulatory DNA that regulate the transcription of target genes by interacting with promoters of target genes. Studies suggested that enhancers can transcribe non-coding RNAs, which was been defined as enhancer RNAs (eRNAs)33. eRNAs were limited to around 500–2000 bp and had shorter half-lives34. Extensive evidences suggested significant roles of eRNAs in tumorigenesis and they involved in various cancer signaling pathways through regulating their target genes. The activation of oncogenic signaling pathways in human cancers often enhanced enhancer activation and production of eRNAs. CELF2 is highly expressed in stomach adenocarcinoma35. APH1A is highly expressed in grade-3 hepatocellular carcinoma (HCC). EN1 is highly expressed in breast cancer (BRAC), and ESR1 can increase eRNA transcription in BRAC36. TAOK1 is associated with overall survival in clear cell renal cell carcinoma22,37. SCRIB was differentially expressed among lung adenocarcinoma patients38. Tumor suppressors-induced eRNAs may implicated in tumor suppression, while oncogene-induced eRNAs can directly promote tumorigenesis. Therefore, eRNAs were closely correlated to malignancy formation and progression.
In our study, we identified 3451 DEGs from 510 THCA samples and 58 normal samples in TCGA cohort, and 362 DE-IRGs (172 downregulated 190 upregulated genes) were extracted. WGCNA was performed to identify THCA-specific immune-related hub genes based on 362 DE-IRGs, and two gene modules (turquoise and grey modules) were obtained. Finally, the grey module was regarded as THCA-specific module. Pearson correlation analysis was used to evaluate immune-related eRNAs. Subsequently, 9-immune-eRNAs prognostic signature (FAAHP1, TP73-AS1, WDFY3-AS2, LINC01184, AL365259.1, TMEM184A, AC007255.1, IQANK1 and AC084375.1) was constructed based on univariate Cox regression analysis and LASSO Cox regression analysis, which was an independent prognostic factor for OS. TP73-AS1, also known as KIAA0495, is abnormally expressed in many cancers39. Previous studies indicated that TP73-AS1 could be a key role in regulating HCC cells proliferation and its expression level was associated with poor prognosis of HCC patients40. Besides, Wang et al. showed TP73-AS1 also interfered the metastasis and proliferation of ovarian cancer41. Of note, we firstly found the expression level of TP73-AS1 is higher in normal thyroid cancer, suggesting that it might have different mechanism in regulating tumor progression comparing other cancers. Increasingly evidences have implicated that lncRNAs participated in the process of cell growth, invasion. Studies showed that overexpressed WDFY3-AS2 suppressed the proliferation, invasion, and epithelial-to-mesenchymal transition (EMT) in ovarian cancer42. Furthermore, WDFY3-AS2 also could promote cisplatin resistance by the expression of miR-139-5p/SDC4 in ovarian cancer, which may provide a promising drug target to drug resistance43. WDFY3-AS2 participated in the development and progression of oesophageal squamous cell carcinoma (ESCC) by regulating miR-2355-5p/SOCS2 axis, which suggested that WDFY3-AS2 might be an underlying predictor and novel therapeutic target for ESCC patients44. Other studies suggested that the potential value of long noncoding RNA WDFY3-AS2 might be novel prognostic biomarker for lung adenocarcinoma (LUAD)45, glioma46, and esophageal cancer (EC)47. AC007255.1, an immune-related prognostic eRNA, was up-regulated in EC tissues and high expression indicated a poorer prognosis. It was closely related to immune response and infiltration levels of immune cells, such as B cell, dendritic cell and neutrophil48. Sui et al. demonstrated that LINC01184 was highly expressed in colorectal cancer, and it could affect the the proliferation and invasion of colorectal cancer cells through the linc01184-miR-331-HER2-p-Akt/ERK1/2 pathway49. Subsequently, the risk score was calculated and separated all patients into high- and low-risk subgroups based on median risk score, and Kaplan–Meier survival curves indicated that the high-risk groups had poorer clinical results than that of the low-risk groups. Univariate and multivariate Cox regression analyses showed the risk score was an independent prognostic factor. Additionally, GO functional analysis suggested that the DEGs between different risk groups were mainly enriched in humoral immune response, immunoglobulin complex, and antigen binding. GSEA showed that humoral immune response, regulation of humoral immune response, and positive regulation of humoral immune response were significantly enriched in low-risk group. To further estimate the TIME of the prognostic signature, ESTIMATE and CIBERSORT were performed to estimate the immune score, stromal score, and tumor purity in THCA sample. More dendritic cells resting and T cells CD4 memory activated were significantly infiltrated in TIME of low-risk group. In addition, patients in the low-risk groups had higher immune score, stromal score and ESTIMATE score than in high-risk group. We applied ssGSEA to assess the immune status of the risk signature, the results suggested that 29 immune-related cells were significantly up-regulated in low-risk group. At the same time, the expression levels of cytokines and immunosuppressor molecules (PD1, PD-L1, PD-L2, CTLA4, TIGIT, TIM-3, BTLA, and LAG3) were significantly higher in low-risk groups, implying more tumor immunogenicity in the low-risk group. The good response of ICIs might be one of the reasons for the good clinical outcome in the low-risk group. Therefore, patients with low-risk scores might be more likely to benefit from ICI treatment.
The present study has some limitations. Our data was retrieved from TCGA public database instead of our own cohort, the predictive power of the prognostic signature should be validated using an external validation cohort, such as GEO cohort. Nevertheless, information on overall survival of THCA patients was unfortunately largely lacking in the GEO databases. Moreover, some basic experiments should be performed to further validate our bioinformatics analysis results in the future. The efficacy of immunotherapy in THCA patients was needed to be validated in large clinical trials.