Lung cancer is one of the main causes of cancer-related death and is responsible for approximately 1.8 million deaths each year (Bray et al., 2018; Hoy et al., 2019). Approximately, 85% of these patients had non–small cell lung cancer (NSCLC), and the rest had small cell lung cancer (SCLC) (Sher et al., 2008; Byers and Rudin, 2015; Denisenko et al., 2018). Lung adenocarcinoma (LUAD) is the most common type of non–small cell lung cancer (NSCLC) and comprises approximately 40% of all lung cancer cases (Denisenko et al., 2018). Despite the improvement in current technology and techniques, the overall survival of LUAD has not been significantly improved, and only a fraction of patients benefited from therapies (Yamanashi et al., 2017; Schenk et al., 2021). Therefore, it is urgent to identify and explore more efficient therapeutic targets to further improve its prognosis.

The Hippo signaling pathway is evolutionarily conserved across higher order vertebrates, and by modulating target genes, it regulates multiple bioprocesses, including cell proliferation, survival, differentiation, and fate determination, as well as organ size and tissue homeostasis (Mohajan et al., 2021). Many of these roles are mediated by the transcriptional effectors Yes-associated protein (YAP) and its paralog transcriptional coactivator with the PDZ-binding motif (TAZ), which direct gene expression by control of a family of sequence-specific transcription factors called TEA DNA-binding proteins (TEAD1–4) that mediate proliferation and pro-survival genes (Dey et al., 2020; Masliantsev et al., 2021; Mohajan et al., 2021). Aberration of the Hippo pathway and YAP/TAZ-TEAD activity was recently shown to be linked to carcinogenesis in lung cancer (Huang et al., 2017; Liu et al., 2017; Gu et al., 2020). Overexpression of YAP/TAZ is associated with the development, progression, and poor prognosis of the disease (Mohajan et al., 2021). Therefore, the Hippo pathway is a novel tumor molecular biomarker and potential therapeutic target for LUAD. As one main component of the Hippo pathway, TEAD4 is a transcriptional enhancer–associated domain (TEAD) family protein (Pobbati and Hong, 2013) that plays biological roles by binding with DNA elements via its specific DNA-binding domains or through interaction with transcription coactivators (i.e., YAP/TAZ) by transactivation domains (Zhou et al., 2017; Gu et al., 2020). More recently, TEAD4 has been demonstrated in tumorigenesis and cancer progression, including cancers of the breast (He et al., 2019; Wu Y et al., 2021), prostate (Chen CL et al., 2021), gastric (Shuai et al., 2020), bladder (Wu et al., 2019; Wang J et al., 2021), thyroid (Zhang et al., 2022), and lungs (Zhou et al., 2017; Gu et al., 2020; Hu et al., 2021; Yan et al., 2022). Previous studies have reported that TEAD4 is upregulated in LUAD and is closely related to disease prognosis (Hu et al., 2021). However, the specific molecular mechanism of TEAD4 regulation on the prognosis of LUAD is not understood.

The aim of the present study was to determine whether TEAD4 could serve as a potential predictor of the prognosis of LUAD. We analyzed TCGA-LUAD samples with high and low TEAD4 expressions, constructed a four-gene prognostic signature based on the TEAD4 differential expression, and determined that TEAD4 was an immune regulation-related predictor of prognosis for LUAD.

LUAD, one of the most frequently diagnosed cancers, is a heterogeneous disease that is commonly triggered by the alteration of key genes including oncogenes and tumor suppressors (de Sousa and Carvalho, 2018; Gavralidis and Gainor, 2020). A better understanding of the risk genes of LUAD and their specific molecular mechanism will facilitate the prevention and management of LUAD. With the rapid development of molecular biological technology and public databases, increasing numbers of biomarkers associated with the prognosis and diagnosis of LUAD have been acknowledged in recent years. However, few factors are of real clinical value.

This current study reveals an immune regulation-related biomarker, TEAD4, for the prognosis prediction and diagnosis of LUAD. TEAD4 is widely studied as a Hippo signaling pathway-related transcription enhancer factor domain family gene, that interacts with YAP/TAZ to act as a transcription factor (Wu Y et al., 2021). This study first showed that the high TEAD4 expression is associated with the prognosis in LUAD patients. Subsequent analysis determined that TEAD4 was an independent prognostic signature in LUAD. In addition, the DEGs related to the TEAD4 differential expression were involved in pathways of the cell cycle, immune response, and metabolism regulation. Increasing evidence suggests that the dysregulation of the immune response has been widely reported to be linked to antitumor immune escape and associated with poor outcomes in cancers (Gerada and Ryan, 2020). Dysfunction of metabolites and their regulators is emerging as a key factor affecting both cancer progression and therapeutic responses (O'Sullivan et al., 2019; Wu Q et al., 2021; Xia et al., 2021). These results all indicate that LUAD patients with different TEAD4 levels may have distinguishing antitumor abilities and further outcomes. TEAD4 has been reported to be a protumor factor in LUAD, including its functions in promoting cancer cell proliferation, migration, and therapy resistance (Zhang et al., 2018; Gu et al., 2020; Hu et al., 2021; Yan et al., 2022). Our conclusions are consistent with these reports. However, the molecular mechanism of this gene in LUAD has not been systematically studied in previous reports. This study conducted a systematic analysis of a large number of samples that, combined with the clinical risk factors for LUAD, revealed more possible mechanisms of its pro-metastatic effects in LUAD and explained more potential reasons for the poor prognosis caused by its high expression at the macro level. This provides possible research directions for further studies on the anticancer function of this gene, such as metabolic regulation and the relationship between macrophage infiltration and TEAD4 disorder. Our result that TEAD4 is downregulated in LUAD seems to conflict with the prognostic result in our study, but it can be explained by the following possibility. First, during tumorigenesis, a large number of genes are changed at the expression level. The general trend is that oncogenes are activated by upregulation, while tumor suppressors are disrupted in the function (Lee and Muller, 2010). These factors all predispose the development of tumor cells. However, tumor cells live in a complex microenvironment composed of a variety of cells, including immune cells, cancer-associated fibroblasts, cancer stem cells, the extracellular matrix, and blood vessels (Arneth, 2019). Tumor growth is jointly regulated by these multiple cells and their secreted factors, and some protumor genes were not upregulated to limit their unlimited growth. Moreover, there is considerable gene expression heterogeneity in tumors among different populations. The results of RNA sequencing represent the overall expression of genes in all cancer patients but not the specific expression of genes in individuals. In summary, the regulatory role of TEAD4 in LUAD still needs to be further explored, which is of great significance for finding novel potential therapeutic targets for LUAD.

Among the 102 DEGs in TEAD4-high expression subtypes, four genes including CPS1, ANLN, RHOV, and KRT6A, were identified to be independent prognostic signatures after univariate Cox regression, LASSO regression, and multivariate Cox regression. The four genes were all positively correlated with the TEAD4 expression in LUAD, indicating that they were TEAD4-related signatures. The enhanced expression of this four-gene signature represents the populations of high risk. However, this was caused by the high TEAD4 expression. In other words, the high expression of TEAD4 led to a poor outcome partly by improving the four-gene signature expression. This study highlighted the relationship between the overall survival and immune microenvironment estimation of patients and the elevated expression of these genes and indicated that the high expression of TEAD4 predicted the poor outcome and the potential immunotherapeutic resistance by improving the four-gene signature expression. In addition, similar to TEAD4, three of these genes (ANLN, KRT6A, and RHOV) were associated with poor outcomes in LUAD, which were all recognized as oncogenic genes. For instance, ANLN is a well-known oncogene that promotes carcinogenesis and therapeutic resistance in multiple types of cancers, such as LUAD (Long et al., 2018; Xu et al., 2019; Deng et al., 2021), oral cancer (Wang B et al., 2021), colorectal cancer (Liu et al., 2022), breast cancer (Wang et al., 2020; Maryam and Chin, 2021), pancreatic cancer (Wang et al., 2019), and head and neck squamous cell carcinoma (Guo et al., 2021). RHOV has been widely studied to promote LUAD cell growth, metastasis, and therapeutic resistance (Chen H et al., 2021; Zhang et al., 2021). In addition, KRT6A has been shown to participate in tumor proliferation, invasion, EMT, and cancer stem cell transformation in lung cancer (Yang et al., 2020; Zhou J et al., 2021; Che et al., 2021). CPS1 has been reported to be an oncogene that is upregulated and has prognostic significance in LUAD (Wu et al., 2020). Our study here found that CPS1, ANLN, RHOV, and KRT6A were TEAD4-related independent prognostic signatures in LUAD. This finding indicates that the four genes were regulated by TEAD4 or the Hippo pathway, which provides an innovative theoretical basis for further research on the regulatory mechanism of these genes. It also provides more possibilities for studying the anticancer mechanism of these genes. Additionally, our study is the first to combine these four oncogenic genes and divide the LUAD sample into two risk subgroups according to the risk score, calculated by the combination of the expressions of the four genes and the survival of patients. Compared with previous studies on the four genes, the present study focuses more on the analysis of the common prognostic value of the four genes in combination with clinical risk factors. These four genes were innovatively identified as independent risk factors for LUAD to predict prognosis in conjunction with other clinical risk factors and provide a new theoretical basis for the choice of individual treatment for patients. Finally, through systematic analysis of data from different databases, the consistent results confirm the prognostic value of the four-gene signature and further highlight the non-negligible role of these genes in human cancer.

The TME estimation of the two risk subtypes suggests that the high-risk group has higher stromal, immune, and ESTIMATE scores, as well as a lower score of tumor purity. The proportions of plasma cells, resting memory CD4 T cells, monocytes, resting dendritic cells, and resting mast cells were lower in the high-risk group, while the proportions of CD8 T cells, activated memory CD4 T cells, resting NK cells, M0 macrophages, M1 macrophages, activated dendritic cells, activated mast cells, and neutrophils were higher. Tumor-associated macrophages (TAMs) are among the most abundant immune cells in the TME and act to enhance tumor progression and metastasis (Mantovani et al., 2017; Cassetta and Pollard, 2018; Lopez-Yrigoyen et al., 2021). High infiltration of TAMs is associated with poor prognosis in several types of cancer, such as breast cancer, ovarian cancer, bladder cancer, and NSCLC (Yin et al., 2017; Zhao et al., 2017; Kowal et al., 2019; Lopez-Yrigoyen et al., 2021). The microenvironmental stimuli and signals that encounter each specific tissue always induce macrophage polarization. According to the specific inducers, two major macrophage subpopulations, classically activated or inflammatory (M1) and alternatively activated or anti-inflammatory (M2) macrophages, have been identified (M0 macrophages are naïve macrophages without polarization) (Miao et al., 2017; Huang et al., 2019; Kang et al., 2020). Functionally, M1 macrophages have robust antimicrobial and antitumoral activity, by removing pathogens during infection (Shapouri-Moghaddam et al., 2018), while M2 macrophages participate in angiogenesis, immunoregulation, tumor formation, and progression (Shapouri-Moghaddam et al., 2018). The different levels of infiltration of immune cells directly determine the different prognoses of patients.

As another indicator of immune escape, MHC-I and MHC-II molecules were found to be lower in the high-risk subtype than in the low-risk group. Degrading MHC-I is always a cause of immune evasion, a major obstacle for cancer therapy, which has been implicated in resistance to immune checkpoint blockade (ICB) therapy (McGranahan et al., 2017; Rodig et al., 2018; Burr et al., 2019; Yamamoto et al., 2020a; Yamamoto et al., 2020b; Zhou Y et al., 2021; Dhatchinamoorthy et al., 2021). During the process of immune evasion, MHC-I downregulation is one major mechanism to avoid antitumor immunity by reducing recognition by cytotoxic CD8⁺ T cells, often correlating with poor prognosis (Cornel et al., 2020). In addition to MHC-I and MHC-II, an antigen-presenting complex traditionally associated with professional antigen-presenting cells (APCs) is critical in antitumor immunity (Axelrod et al., 2019). Tumor-specific MHC-II is reported to be associated with superior prognosis, allowing recognition of tumor cells by the immune system, thus playing a role in immunotherapy and improving the response to ICB therapy (Mortara et al., 2006; Forero et al., 2016; Johnson et al., 2016). The downregulated MHC-I and MHC-II molecules in the high-risk subtypes indicate the potential strong immune escape and ICB therapy resistance of the patients in this group.

Moreover, immune checkpoint genes were also detected, and the results showed that CD274 (PD-L1) and LAG3 (CD223) were highly expressed in the high-risk subtype. PD-L1 (Programmed death-ligand 1) is expressed on several types of tumor cells, mediating the tumor-induced immune suppression (immune checkpoint) by binding with the receptor PD-1 (programmed cell death protein 1), which is highly expressed in activated T cells, B cells, dendritic cells, and natural killer cells (Dermani et al., 2019). The binding of PD-L1 to PD-1 on T cells results in the inhibition of cancer cells by destruction by T cells, thus promoting immune escape (Gou et al., 2020). Therefore, PD-L1 or PD-1 monoclonal antibodies have been used for cancer treatment (Bagchi et al., 2021; Carlino et al., 2021; Doroshow et al., 2021). A higher level of PD-L1 predicts a worse outcome in patients. In addition, LAG3 (lymphocyte activation gene 3, CD223) is another kind of inhibitory receptor (IRs) that has been reported to play a negative regulatory role in cancer immunology by interacting with its ligands (Wang M et al., 2021). LAG3 expression is also shown to be positively associated with CD274 (PD-L1) (Wang M et al., 2021). TEAD4 is known as a transcription factor associated with resistance to different therapeutic approaches (Jiao et al., 2017; Yu et al., 2021; Yan et al., 2022). Unfortunately, these studies were not confirmed in patients with LUAD. Our study systematically analyzed the TME in patients with LUAD and demonstrated that a high expression of TEAD4 is associated with a poor anti-tumor immune response, with evidence of a lower immune score and HLA family components and higher levels of immune checkpoint genes in the high-risk subgroups based on the high TEAD4 expression. This not only reinforces previous research but also provides new insights into the mechanisms of this gene involved in therapy resistance. Finally, DEGs between the two risk groups were identified, and the subsequent GO and GSEA showed that cell division and cell cycle regulation-related genes were enriched in the high-risk group, while immune response- and metabolism-related genes were enriched in the low-risk group. This is consistent with TEAD4-high expression-related DEGs, as well as the TME result in the Cox model, further confirming the conclusion in this study.

According to these DEGs, the forecasted adjuvant small-molecule drugs for the high-risk subtype are perturbations targeting CFTR, PLA2G1B, CDC25A, CHIA, TTK, and OXM1 by the moa of the CDK inhibitor, CFTR channel agonist, cytochrome p450 inhibitor, glucocorticoid receptor agonist, cyclooxygenase inhibitor, or NFkB pathway inhibitor. Targeting these pathways may be an efficient therapeutic strategy for patients with high levels of TEAD4. A total of 28 potential small-molecule drugs were predicted based on the specific differentially expressed genes in these populations. This not only provided a novel solution to the low survival of patients with LUAD but also laid a theoretical foundation for further drug research and development.

However, there are several limitations to the present study that should be stated. First, despite the fact that bioinformatic technology is powerful in efficiently understanding biological functions, the underlying mechanisms of these genes in LUAD still need further cellular explorations. Moreover, clinical tissues and paired adjacent normal tissues should be collected to further detect the protein expression level of TEAD4, as well as the related cell cycle and immune response genes in LUAD. Second, as this is a retrospective study, missing data and selection biases were inevitable, and the statistical power might be low. Therefore, further studies with a large sample size are warranted to increase the statistical power. Finally, due to lack of data about immunotherapies, the relationship between the TEAD4 expression and ICB therapy response cannot be investigated. More clinical and demographic characteristics of LUAD patients need to be included in further studies.

In summary, our results suggest that TEAD4 is a novel molecular biomarker for diagnosis and prognosis, predicting overall survival and immune microenvironment estimation in LUAD. However, large prospective studies are warranted, and further experimental validation should be performed to prove the prognostic value of this gene in LUAD.