TAO is a complex autoimmune disease, and its precise molecular mechanisms have not been completely elucidated; moreover, effective treatment options for TAO are limited . Therefore, it is worth exploring novel biomarkers and therapeutic targets to improve the prognosis of TAO patients. With the application of bioinformatics, some studies have been performed to elucidate the pathogenesis of TAO. However, overreliance on public datasets and sequencing technology may destabilize results. In addition, most studies to date have focused on transcriptional regulation in a single tissue and the similarity and difference in transcription regulation among different lesion tissues are relatively unknown. In the present study, considering the tissue-specific effects, two public gene expression profiles containing different lesion tissues (lacrimal gland and orbital adipose/connective tissue) from TAO patients (GSE105149 and GSE185952) were used to screen and investigate key genes and pathways to examine the biological and clinical implications of TAO. Additionally, collected the serum samples of 50 TAO patients and 20 healthy controls to verify our findings through the ELISA. The obtained genes and pathways may provide novel insight into the underlying mechanisms and may reveal a new strategy for the diagnosis and treatment of TAO.
In this study, we separately analyzed the differentially expressed genes and pathways in the GSE105149 and GSE185952 datasets. To investigate the potential biological association between different lesion tissues in TAO, the results of functional enrichment analysis were compared, and three overlapping pathways (neuroactive ligand-receptor interaction, IL-17 signaling pathway, and TNF signaling pathway) were found between the two datasets. Therefore, it is logical to speculate that these three overlapping pathways may play key roles in TAO progression and development; other differential pathways might be influenced by tissue or organ specificity. In fact, consistent with our findings, several studies have confirmed that certain biomarkers and pathways, including cytokines, pro-inflammatory markers, and inflammation-related pathways, appear to be less specific to TAO . Additionally, we sought to explain why the three overlapping pathways are not influenced by tissue specificity. Autoimmune responses are considered to be central in the development and progression of TAO. The IL-17 and TNF signaling pathways are inflammatory pathways associated with TAO [26,27,28]. Furthermore, preliminary studies have found that blocking IL or TNF-α pathways may constitute a systemic treatment strategy for TAO [29, 30]. This evidence suggests that the two inflammation-related pathways might be involved in the occurrence and development of TAO, which may account for the lack of influence of tissue specificity. Similarly, growing evidence shows that some neuromodulation-related pathways may be involved in TAO [31, 32]. Thus, inflammatory mechanisms and neuronal function may play important roles in the different lesion tissues of TAO. However, the exact mechanisms underlying these pathways remain unclear.
In the present study, a set of 52 common DEGs from the GSE105149 and GSE185952 datasets were identified, including 35 up-regulated genes and 17 down-regulated genes. To further explore the biological function of these 52 common DEGs, functional enrichment analysis was performed; long-term depression, the VEGF signaling pathway, the IL-17 signaling pathway, the Ras signaling pathway and inflammatory mediator regulation of transient receptor potential (TRP) channels were the top 5 pathways. The IL-17 and TNF signaling pathways showed enrichment, while the neuroactive ligand-receptor interaction pathway was not directly included. Other neuromodulatory signaling pathways were also enriched, such as serotonergic synapse and glutamatergic synapse pathways, which was indirectly verified by our findings. Surprisingly, long-term depression containing PLA2G4D and PRKCG genes was the top pathway, suggesting that this pathway and these genes may have novel potential biological implications in TAO. PLA2G4D, a member of the phospholipase A2 (PLA2) family, promotes production of lipid mediators and regulates normal physiology and disease pathogenesis in many organ systems . One previous study confirmed that another three family members, PLA2G4A, PLA2G4B, and PLA2G4C, are strongly associated with mental illness in the Chinese population . These findings add support to the results of our previous study using diffusion tensor imaging that TAO patients develop disruption of the structural brain network connectome, which may be associated with clinical-psychiatric dysfunction in TAO . Previous studies have also demonstrated that TAO patients are more prone towards emotional disorders, such as anxiety and depression [36, 37]. Nevertheless, the specific relationship between TAO and depression remains to be investigated.
In the present study, six key genes (PRKCG, OSM, CSF3R, LRRTM1, CXCL6, and DPP4) among 52 common DEGs were analyzed in the PPI network based on interactions. PRKCG encodes protein kinase C gamma (PKC gamma), a member of the serine/threonine kinase family, which is abundantly expressed in neural tissue . Studies have shown that PKC dysregulation is associated with major depression [39, 40]. LRRTM1, an imprinted gene that affects neuronal differentiation and connectivity, correlates with schizophrenia/schizoaffective disorder . However, the precise roles of PRKCG and LRRTM1 in TAO are unknown, and we suspected they may be involved in the mechanisms of the impaired emotional regulation and higher tendency to depression or anxiety occurring in TAO patients [3, 35]. In addition, four other key genes, CSF3R, OSM, CXCL6, and DPP4, showed positive associations with inflammatory processes in TAO. For example, CSF3R is the receptor for colony stimulating factor 3, a cytokine that controls expansion and differentiation of neutrophils. Mutations in CSF3R have been linked to chronic neutrophilic leukemia . To our knowledge, no study to date has analyzed the effects of CSF3R in TAO. OSM is a proinflammatory cytokine that belongs to the interleukin-6 (IL-6) family. It is produced mainly by activated macrophages, neutrophils and mast cells . Chen et al. discovered that IL-6 is overexpressed in the orbital tissue of patients with TAO, promoting proliferation and differentiation of B cells, and activation of T helper 1 (Th1) cells, releasing a variety of cytokines . However, a possible role for OSM in TAO has not been reported. CXCL6, a member of the CXC chemokine family, is highly expressed in patients with diabetic nephropathy (DN) . In an in vitro study, overexpression of CXCL6 significantly up-regulated expression of proinflammatory and profibrotic cytokines . Of note, Antonelli et al. reported that interferon-gamma (IFN-gamma) and TNF-alpha stimulate release of chemokines, particularly CXCL9, CXCL10 and CXCL11, and induce recruitment of activated T cells to maintain inflammation in TAO . Whether CXCL6 has similar roles in TAO remains unknown. In fact, proinflammatory cytokines have been extensively studied in recent years in relation to the pathogenesis of TAO [48, 49], and our functional enrichment results indicated that OSM, CXCL6 and CSF3R genes are all closely related to the cytokine-cytokine receptor interaction pathway. It will be helpful to further clarify their specific roles in the inflammatory process. DPP4, also known as CD26, is a cell surface glycoprotein receptor that plays an important role in regulating T-cell activation. Previous studies have indicated that IFN-gamma and IL-12 can facilitate expression of CD26 on CD4 + T cells, thereby promoting Th1-like cytokine production . It is well known that in the early stage of TAO production of pro-inflammatory cytokines by Th1 cells can enhance fibroblast proliferation and glycosaminoglycan production. Therefore, we speculate that DPP4 might be involved in the pathogenesis of TAO.
In the current study, PRKCG, OSM, CSF3R, LRRTM1, and CXCL6 were up-regulated and DPP4 was down-regulated in TAO. Considering the limitations of public datasets and bioinformatics, we verified the results using clinical samples from our center. As no commercial ELISA kit was available for testing serum LRRTM1, only PRKCG, OSM, CSF3R, CXCL6 and DPP4 protein levels were detected via ELISA in 50 TAO patients and 20 HCs. Overall, protein expression levels of PRKCG, OSM, and CSF3R were significantly higher in TAO patients, consistent with the bioinformatics results. Similarly, protein expression level of CXCL6 exhibited an up-regulated tendency in TAO, though the difference was not significant. This may be due to individual variation in CXCL6 expression among TAO patients. Interestingly, the mRNA and protein expression levels of DPP4 seem to be contradictory, as DPP4 protein expression was upregulated in TAO patients, but mRNA expression was decreased, which may be ascribed to the non-strict linear relationship between mRNA and protein as well as more complex dependencies. For example, different regulatory mechanisms (RNA modifications, RNA-RNA interactions, synthesis and degradation rates) might explain the different results in DPP4 mRNA and protein expression levels , and more samples are needed to explore this possibility. According to multivariable regression, we further found that the expression levels of DPP4 and PRKCG were significantly associated with TAO occurrence. The area under the ROC curve illustrated that these two genes exhibit good diagnostic value for TAO. Additionally, CSF3R is a promising candidate for distinguishing between active TAO patients and inactive ones. Therefore, these hub genes may play important roles in the pathogenesis of inflammatory processes, with great potential as reliable and robust blood diagnostic biomarkers and therapeutic targets for TAO. In summary, the novelty of this study mainly includes the following aspects: First, to the best of our knowledge, the six genes (PRKCG, OSM, DPP4, LRRTM1, CXCL6, and CSF3R) identified are for the first time reported to be differentially expressed in various samples (lacrimal gland and orbital adipose/connective tissue) of TAO patients and defined as hub genes for TAO; Second, we first report the interactions of these hub genes in TAO patients, and these hub genes might be associated with disease progression and clinical characteristics of TAO; Third, bioinformatics analysis combined with molecular biology was used to explore potential genes and pathways in TAO patients, and the obtained genes and pathways provide novel insight into the underlying mechanisms and may reveal a new strategy for diagnosis and treatment of TAO.
However, there are still some limitations and prospects in our study. First, although the latest datasets for different lesion tissues of TAO were selected, the sample size was small, and the results need to be confirmed in larger samples; Second, as the current work was based on omics-research and focused on marker screening, few functional studies of these biomarkers were performed. Therefore, it is necessary to further explore the exact mechanisms of TAO influenced by these biomarkers; Third, multiple studies have confirmed the value of long noncoding RNAs (lncRNAs) and microRNAs (miRNAs) in disease development, prediction and treatment [52, 53], and several lncRNA- and miRNA-disease associations might be identified by novel state-of-the-art computational models [54,55,56]. However, our study did not cover these aspects. Elaborating on such connections will be a good direction for future research.