TWF1 induces autophagy and accelerates malignant phenotype in lung adenocarcinoma via inhibiting the cAMP signaling pathway

Many studies have shown that the actin cytoskeleton plays an essential role in the initiation and progression of cancer. As an actin‐binding protein, Twinfilin1 (TWF1) plays an important role in regulating cytoskeleton‐related functions. However, little is known about the expression and function of TWF1 in human tumors. The present study aimed to investigate the functional roles and the underlying molecular mechanisms of TWF1 in human lung adenocarcinoma (LUAD). By using bioinformatics databases and tumor tissues, TWF1 expression was found to be higher in LUAD tissues than in adjacent tissues and poor survival was predicted in patients with LUAD. In vitro and in vivo assays indicated that downregulation of TWF1 expression suppressed LUAD cells invasion and migration. Further studies revealed that TWF1 interacted with p62 and was involved in the regulation of autophagy. The molecular mechanisms underlying TWF1 were investigated by RNA‐seq analysis and a series of functional experiments. The results showed that downregulation of TWF1 suppressed LUAD progression through the cAMP signaling pathway. Therefore, overexpression of TWF1 in LUAD promoted migration, invasion, and autophagy through the cAMP signaling pathway.


| INTRODUCTION
The incidence of lung adenocarcinoma (LUAD) has increased steadily over the past 2 decades, and it is currently the most common type of non-small cell lung cancer (NSCLC) in the United States, Europe, and East Asia. 1,2 Although the current treatment of LUAD is constantly being updated, the poor prognosis remains a challenge due to the high incidence of recurrence and metastasis. 3 This disappointing fact could be largely attributed to the lack of specific postoperative monitoring biomarkers and effective therapeutic targets. 4 Therefore, the search for new biomarkers and effective targets is the key to improving patient outcomes.
Autophagy is an evolutionarily conserved cellular process that enables the orderly degradation and recycling of cellular components, which are critical for maintaining cellular homeostasis. 5 In cancer development, autophagy plays a dichotomous role in suppressing the early stages of tumorigenesis through autophagy-mediated cell killing but favoring advanced tumor progression and conferring resistance to therapies. 6 Targeting modulators of autophagy signaling is a promising strategy for the development of anticancer drugs. Several studies have provided compelling evidence that autophagy inducers and inhibitors have shown promising results in preclinical research. 7,8 In particular, autophagy inhibitors, such as chloroquine or hydrochloroquine (HCQ), have shown promising efficacy in several tumors, especially in combination with chemotherapy and targeted-therapy, and they are undergoing clinical trials. 9 Some researchers have linked autophagy to the inhibition of LUAD progression. 10 Other studies have shown that autophagy could facilitate LUAD tumorigenesis and development. 11 Therefore, regulation of autophagy activity based on underlying molecular mechanisms and the specific targets of autophagy may be considered as an effective interventional strategy for cancer prevention and therapy.
Twinfilin 1 (TWF1), an actin monomer-binding protein first described in yeast, is an evolutionarily highly conserved protein. 12 Consisting of two homologous domains of actin-depolymerizing factor (ADF), TWF1 is mainly found in the cytoplasm where it binds to actin monomers and regulates cytoskeletal assembly/disassembly. 13 Previous pan-cancer analysis showed that TWF1 was upregulated in various solid tumors and suggested poor prognosis, including LUAD, mesothelioma, cervical cancer, and pancreatic cancer. 14 Currently, few studies could be found on TWF1 in tumors and they generally focused on the tumor-promoting role of TWF1 by regulating cytoskeletal structure, 15 including chemoresistance in breast cancer, hepatocellular carcinoma (HCC), and pancreatic cancer [16][17][18] and epithelial-mesenchymal transition (EMT) in clear cell renal cell carcinoma. 19 In addition, TWF1 has been implicated as a key regulator of cisplatin resistance and EMT in NSCLC. 20 However, the role of TWF1 in LUAD and the molecular mechanism involved remain unclear. In addition, recent evidence supports that the cytoskeleton plays an important role in autophagy in yeast and mammalian cells, affecting organelle recycling and degradation. 21 Relatively little is known about the role of the actin cytoskeleton and autophagy in tumor. Clearly, elucidating the function and regulatory mechanism of TWF1 in LUAD, particularly its role in autophagy, may provide a new therapeutic strategy for cancer treatment.
In this study, by using database exploration of The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO), TWF1 was confirmed to be highly expressed in LUAD and strongly associated with poor prognosis of patients with LUAD. Downregulation of TWF1 expression inhibited the proliferation and metastasis of LUAD cells in vitro. In addition, TWF1 could promote malignant progression of LUAD cells by activating autophagy through inhibition of the cAMP/PKA/CREB1 signaling pathway. Finally, a xenograft model showed that TWF1 induced LUAD progression in vivo. In conclusion, this study may provide a new strategy for targeting TWF1 in the treatment of patients with LUAD.

| Patients and tissue specimens
Fresh human LUAD and adjacent nonmalignant tissue samples were obtained from the Tianjin Medical University Cancer Institute and Hospital between January 2021 and December 2022. None of the patients had undergone adjuvant therapy such as radiotherapy or chemotherapy before surgical resection. All tissue samples were verified as LUAD or adjacent normal tissue by two independent pathologists, and the LUAD samples were staged based on the 8th edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual. The study was consistent with the ethical guidelines of the Helsinki Declaration and approved by the Ethics Committee.

| Cell culture
A549 and H1299 cells were purchased from American Type Culture Collection. The cell lines were cultivated in 1640 medium and supplemented with 1% penicillin/streptomycin (PS; HyClone) and 10% fetal bovine serum (FBS; PAN-Seratech). The incubating temperature was 37°C, with 5% CO 2 .

| Transcriptome sequencing analysis
Transcriptome sequencing was based on the Illumina sequencing platform to analyze the gene expression of A549 sh-TWF1 cells and sh-Ctrl cells. DEGs were filtered by log2 (fold change) ≥1 and p-value ≤ .05. ClusterProfiler was used for gene enrichment analysis of gene ontology and KEGG, and p-value ≤ .05 was used as the enrichment cut off.

| Autophagic flux
Cells were transfected with the stubRFP-sensGFP-LC3 lentivirus purchased from GENE-CHEM following the manufacturer's protocol. After transfection with mRFP-GFP-LC3, autophagosomes were labeled yellow (mRFP and GFP) and autolysosomes were labeled red (mRFP only). Then, the cells were visualized using fluorescence microscopy.

| Animal models
For animal experiments, male BALB/c nude mice (4 weeks old) were purchased from SPF Biotechnology. tumor cells (5 × 10 6 ) prepared at a volume of 100 μL were injected into the subcutaneous tissue of each mouse by a 1 mL injector (n = 20 in the A549 sh-TWF1-A and sh-Ctrl groups). The status of mice was observed every 2 days. The tumor volume was checked with a caliper, and the variation of weight of mice was recorded by a scale.

| Colony formation assay
For colony formation assay, 1000 cells in DMEM supplemented with 10% FBS were plated in six-well plates. After 2 weeks of incubation, the surviving colonies were fixed, stained with 0.5% crystal violet, imaged, and counted. The data are presented as the means ± SDs of triplicate dishes in the same experiment.

| TCGA and GEO datasets
The raw data of TCGA (National Cancer Institute) and GEO (NCBI) related to LUAD were downloaded on the official website. Then, the data were normalized by R Studio.
The expression of TWF1 was analyzed in matched tumor and para-tumor of TCGA and GEO databases (GSE116959, GSE32665, GSE43458, and GSE27262). Further analysis of the expression of TWF1 with the prognosis of patients with LUAD was conducted.

| Co-immunoprecipitation
To identify potential TWF1-interacting proteins among Beclin1, Atg5, and P62 at the endogenous level, A549 and H1299 cells were washed with ice-cold PBS three times before being lysed in IP lysis buffer. Then, the lysates were incubated with anti-TWF1/anti-p62/IgG antibody overnight at 4°C. Protein A/G-agarose beads were added for 1 h at RT. The beads were collected and washed with lysis buffer for three times. The precipitated proteins were eluted and denatured in 1 × SDS loading buffer and analyzed by western blotting. The following antibodies were used: anti-TWF1 (1:1000), anti-SQSTM1/p62 (1:1000) from Santa Cruz Biotechnology, and anti-Atg5 (1:1000), anti-Beclin 1 (1:1000) from Cell Signaling Technology.

| Immunohistochemistry
Xenograft tumor tissue sections were deparaffinized in xylene and rehydrated through an ethanol series. Antigen was then retrieved in citrate and treated with 3% hydrogen peroxide to inhibit endogenous peroxide activities for 10 min. Then, the samples were stained using antibodies at room temperature for 30 min and overnight at 4°C. After the samples were washed, tissue microarrays and sections were incubated with secondary antibody for 1 h at room temperature, visualized with 3,3-diaminobenzidine solution (ZSGBBio) treatment, and counterstained with hematoxylin.

| TWF1 is overexpressed in LUAD tissues and associates with poor prognosis in patients with LUAD
Our previous study showed that TWF1 is highly expressed in a verity of tumors and closely associated with poor prognosis. In the present study, publicly available LUAD transcriptome profiling data were collected to analyze the potential relationship between TWF1 expression and LUAD. In addition to TCGA data, GSE116959 (57 LUAD samples and 11 peritumoral normal lung tissues), GSE32665 (87 lung adenocarcinoma and 92 adjacent uninvolved lung tissue tissues, 85 matched pairs), GSE43458 (30 normal lung tissues paired with 30 never-smoked cases), GSE27262 (tumor and adjacent control tissue pairs from 25 patients with LUAD), GSE140797 (seven pairs of lung adenocarcinoma tissues and normal tissues), and GSE7670 (a total of 66 samples, including 27 paired samples from patients who underwent surgery for lung cancer) were included. By using these comprehensive datasets, TWF1 was found to be highly expressed in LUAD in GSE116959, GSE32665, and TCGA database ( Figure 1A-C). Moreover, the TWF1 transcription in TCGA LUAD samples was higher than in normal samples of GTEx combined TCGA database ( Figure 1D). These results were still accurate in 231 pairs of LUAD tissues and matched non-cancer tissues in GSE43458, GSE27262, TCGA, GSE140797, GSE7670, and GSE27262 databases ( Figure 1E-G and Figure S1A-C). The characteristics of 535 patients with primary LUAD, as shown in Tables S1 and S2, were collected from the TCGA database for cases where clinical and gene expression data were available. On the basis of the relative TWF1 expression levels, patients with LUAD were divided into two groups: high-(n = 268) and low-(n = 267) expression groups. The correlations between the expression of TWF1 and the clinicopathological characteristics of LUAD patients were compared. Then, ROC curve was used to investigate whether the expression of TWF1 could separate LUAD tissues from non-tumor tissues. The area under the curve (AUC) of TWF1 was 0.854, indicating that TWF1 expression could be used as a biomarker in the diagnosis of LUAD ( Figure 1H). The Kaplan-Meier analysis was used to evaluate the correlations between TWF1 expression and the overall survival (OS), progression-free survival (PFS), and disease-specific survival (DSS) of patients with LUAD. The results indicated that patients with higher TWF1 expression had worse OS, PFS, and DSS in LUAD (p < .01, Figure 1I-K). Further investigation of the expression in LUAD showed that the protein levels of TWF1 in four pairs of fresh LUAD tissue samples were significantly higher than those in adjacent normal tissue samples ( Figure 1L,M).

| Knockdown of TWF1 expression suppresses LUAD cells migration and proliferation in vitro
A wild-type LUAD cell line, A549, was selected to construct cell lines with stable TWF1 downregulation (sh-TWF1-A, sh-TWF1-B) by transfection with lentiviral plasmids to further investigate the role of TWF1 in LUAD tumor progression. Meanwhile, sh-Ctrl control cells were also constructed. The efficiencies of TWF1 deletion were confirmed by western blotting  and RT-PCR (Figure 2A). Similarly, the expression of the TWF1 at the protein and mRNA levels were suppressed in the reduced expression group H1299 cells (sh-TWF1-A) ( Figure 2B). Next, the biological functions of TWF1 in LUAD cells migration were further investigated. Migration and invasion assays confirmed that the downregulation of TWF1 expression suppressed the migration and invasion abilities of A549 and H1299 cells ( Figure 2C-F). The wound-healing assay also confirmed that TWF1 promoted the ability of non-directional migration in A549 cells ( Figure 2G). Therefore, it was suggested that TWF1 gene was significantly associated with the migrating ability of A549 and H1299 cells. In addition, the colony-formation ability of A549 and H1299 cells in the TWF1 downregulation group was significantly suppressed as revealed by the clonogenic assays ( Figure 2H,I). Thus, it was suggested that TWF1 gene was significantly associated with the proliferation ability of A549 and H1299 cells. This series of results all suggested that TWF1 can promote migration and proliferation of LUAD cells.

| TWF1 induces autophagy of LUAD cells via interacting with p62
A complex crosstalk has been reported between cancer metastasis-related and autophagy-correlated signaling pathways. In general, autophagy activation and inhibition may promote or hinder tumor metastasis depending on the types and stages of tumor. 22,23 Therefore, inhibiting tumor progression by targeting autophagy may be a novel strategy. RNA-seq analysis of A549 cells, which were stably expressing TWF1 knockdown (sh-TWF1-A), and control cells were performed to further investigate the mechanism of TWF1 affecting the motility and invasion of LUAD cells. Analysis of the sequencing data results found that TWF1 was closely related to autophagy-associated proteins ( Figure 3A). Some of the top differential genes were also related to autophagy function ( Figure 3B). Therefore, we speculated that TWF1 was closely related to autophagy. Afterwards, transmission electron microscope (TEM) results showed that the number of autophagosomes and autophalysosomes decreased in the TWF1 downregulation group ( Figure 3C). The mRFP-GFP-LC3B double-fluorescence system was used to enhance the understanding on the effect of TWF1 on autophagy. The results showed that the number of yellow dots (autophagosomes) and redonly dots (autolysomes) decreased in TWF1 knockdown A549 cells ( Figure 3D). The expression of autophagyrelated proteins was also detected via western blotting. Reduced TWF1 expression resulted in decreased levels of LC3B-II, Beclin 1, p62, and Atg5 in A549 and H1299 cells ( Figure 3E,F). We further treated A549 cells with rapamycin to activate autophagy. Interestingly, in TWF1 knockdown and control cells, the expression levels of autophagy-related proteins were reversed ( Figure 3G). Next, to probe into the potential molecular mechanism by which TWF1 regulates autophagy in lung cancer cells, we performed Co-IP assay to identify potential TWF1-interacting proteins in A549 and H1299 cells. The Co-IP experiments using endogenous proteins demonstrated that TWF1 and p62 could interact with each other in A549 and H1299 cells ( Figure 3H). These data collectively suggested that TWF1 regulated autophagy via interaction with p62 in LUAD cells.

| TWF1 promotes autophagy, invasion, and migration through cAMP signaling pathway in LUAD
RNA-seq analysis of A549 cells with stably downregulated TWF1 expression was performed to further explore the downstream molecular mechanism of TWF1. With the threshold of a p-value < .05 and a |log2 FC | ≥ 1, a total of 1148 DEGs were detected, including 947 upregulated and 201 downregulated DEGs. Pathway enrichment was then performed on the basis of these DEGs. Notably, in addition to genes in pathways involved in cancer, genes in the cAMP signaling pathway were significantly enriched ( Figure 4A). Most of the genes related to the cAMP signaling pathway were also upregulated, suggesting that the cAMP signaling pathway may be activated after TWF1 downregulation ( Figure 4B). Western blotting revealed that downregulated TWF1 expression promoted the protein levels of p-PKA and p-CREB1 in A549 and H1299 cells ( Figure 4C,D). TWF1 downregulated A549 cells were treated with the cAMP inhibitor H89-2HCI to investigate the relationship between the cAMP signaling pathway and autophagy, invasion, and migration. Fewer yellow dots and red-only dots were observed in the TWF1-downregulated A549 cells and this trend was reversed by the cAMP inhibitor H89-2HCI ( Figure 5A). Western blotting further showed that the H89-2HCI-mediated inhibition of p-PKA significantly increased the levels of autophagy-related proteins LC3B-II, p62, and Atg5 in the TWF1 downexpressing A549 cells ( Figure 5B). The invasion and migration phenotypes in the TWF1 downregulated cells and controls treated with or without H89-2HCI were investigated. As expected, suppression of activated p-PKA and p-CREB1 successfully reversed the malignant behavior of A549 cells caused by the downregulation of TWF1 ( Figure 5C-E).  10×). (E) Woundhealing assay comparing the migration distance between the sh-Ctrl and sh-TWF1 groups of A549 cells. p-, phosphorylation protein content; t-, total protein content. Data were presented as mean ± SEM. n = 3. *p < .05, **p < .01, ***p < .001.

| TWF1 promotes LUAD cell tumorigenesis in vivo
Nude mice were injected subcutaneously with A549 sh-Ctrl cells and A549 sh-TWF1-A cells (5 × 10 6 cells/mouse) to further validate the oncogenic activity of TWF1 in vivo. Tumor volumes were measured every other day from the second week after the injection. All mice were sacrificed at the end of the fourth week and the primary tumors are shown in Figure 6A. Tumor weights and sizes were significantly lower in the sh-TWF1-A group than in the sh-Ctrl group ( Figure 6B,C). Furthermore, the TWF1 expression in the two groups was determined by immunohistochemistry ( Figure 6D). As shown in Figure 6E,F, the TWF1 downregulation group had a significantly lower autophagy level than the control group. Then, indicators related to the cAMP signaling pathway were examined by western blotting (Figure 6E,F). In conclusion, high expression of TWF1 clearly promoted LUAD proliferation and autophagy in vivo.

| DISCUSSION
Lung cancer is known as one of the most common cancers worldwide, with an overall high mortality rate. 24 On the basis of histopathological characteristics, lung cancer could be classified into NSCLC and small cell lung cancer (SCLC). NSCLC is further divided into three major subtypes, of which LUAD is the most common subtype, accounting for 50%, followed by squamous cell carcinoma (SCC) and large cell lung cancer (LCLC), accounting for 40% and 10%, respectively. 25 Despite the immune-based therapies and the emergence of targeted therapies, which have significantly improved the prognosis of selected patients, the average 5-year survival rate of LUAD ranges is between 15% and 17%. 26 The low survival rate could be explained by the lack of reliable biomarkers, leading to delays in treatment. 27 Therefore, the research for new biomarkers, especially in the early stages of the disease, is of great importance for the appropriate diagnosis and treatment of LUAD.
Tumor recurrence and metastasis are the main causes of high mortality in patients with LUAD. 28 Investigation of the key molecules and biomarkers that regulate LUAD recurrence and metastasis is an important prerequisite for the development of effective therapeutic drugs. In the present study, the expression of TWF1 was found to be upregulated in LUAD tissue samples and associated with poor prognosis in patients with LUAD through analysis of TCGA and GEO datasets and western blotting. Furthermore, the AUC of TWF1 was 0.854, indicating that TWF1 expression could be used as a biomarker for the diagnosis of LUAD. These results indicated that TWF1 serves as a key factor mediating the development and progression of LUAD. They were consistent with Li's findings 29 and further supported the possibility of targeting TWF1 for LUAD treatment.
This study showed that downregulation of TWF1 expression could suppress the migration and invasion of LUAD cells. In addition, TWF1 promoted the proliferation of LUAD cells. Notably, an in vivo experiment further supported these phenomena. Therefore, the oncogenic functions of TWF1 in LUAD cells were preliminarily determined in vivo and in vitro. On the basis of a series of experiments, this study hypothesized that TWF1 may act as a potential target in patients with LUAD.
Notably, the enrichment between TWF1-knockdown cells and control cells highlighted a highly differentially expressed gene, the human proprotein convertase subtilisin/kexin 9 (PCSK9). It was firstly reported in 2003 30 and it has been implicated in the regulation of cholesterol homeostasis. 31 Emerging data showed that PCSK9 was abnormally expressed in various malignancies and associated with adverse clinical outcomes. 32 In addition, PCSK9 was involved in the development of autophagy, as discovered in ischemic myocardial cells from a report in 2018. 33 Autophagy is generally considered to be a mechanism that allows cancer cells to maintain cellular homeostasis under either normal or stressful conditions. 34 Accumulating evidence has indicated that autophagy plays a complex role in cancer cell motility and invasion during tumor metastasis. 23 Depending on the tumor microenvironment, autophagy has been reported to play pro-and anti-metastasis roles. 35 In the present research, whole-transcriptome sequencing and bioinformatics analysis showed the expression difference of several autophagy-related genes between TWF1-knockdown cells and control cells. This finding suggested that TWF1 downregulation may affect the autophagy in LUAD cells. Western blotting confirmed the following results: TWF1 downregulation inhibited the levels of p62, Beclin 1, Atg5, and LC3B. Furthermore, the expression of autophagy-related genes were upregulated upon exogenous addition of rapamycin (an autophagy activator) in TWF1 downexpressing LUAD cells. P62 is a selective autophagy receptor with numerous interacting partners and involved in tumorigenesis and development. Kong et al. suggested that DDX5 promoted autophagy of HCC cells by binding to p62 to inhibit tumor proliferation. 36 In head and neck squamous cell carcinoma, p62 interacted with FN1 to accelerate the degradation of FN1, thereby impairing the EMT. 37 In our study, CO-IP assay was performed to identify the potential interacting proteins of TWF1. The results illustrated that p62 acted as a direct interacting partner of TWF1, involved in the regulation of autophagy in LUAD cells. Thus, autophagy may have contributed to TWF1-mediated LUAD cell proliferation and metastasis, which requires further assays to be validated.
Another intriguing phenomenon was that the DEGs induced by the downregulation of TWF1 expression were mainly enriched in the cAMP/PKA/CREB1 signaling pathway. As an important intracellular signaling molecule, cAMP has been reported to be involved in a wide range of physiological processes, such as metabolism, cell proliferation, differentiation and apoptosis, gene expression, and cell death. 38 Recent studies have shown that the cAMP signaling pathway and its downstream effectors have paradoxical effects on tumor development depending on the type of tumor and its microenvironment 39 In glioblastoma, cAMP/PKA signaling pathway has been reported to inhibit tumor cell migration and invasion by downregulating the FAK/AKT pathway. 40 Shen et al demonstrated that diosgenin inhibited malignant phenotypic transformation of colorectal cancer (CRC) by interfering with the cAMP/ PKA/CREB signaling pathway, suggesting that the cAMP signaling pathway may serve as a tumor-suppressive mechanism in CRC. 41 In addition, radiation-induced apoptosis could be enhanced by the cAMP signaling system through suppressing the activation of ataxiatelangiectasia mutated (ATM) protein kinase in NSCLC cells. 42 Consistently, the data presented here showed that downexpression of TWF1 in LUAD cells upregulated the phosphorylation of CREB1 and PKA. A series of rescue experiments was performed to further clarify this point. The cAMP/PKA/CREB1 signaling pathway was inhibited again by exogenous addition of the PKA inhibitor H89-2HCI in TWF1 downregulated LUAD cells. More importantly, in vitro functional experiments confirmed that the phosphorylation state of the cAMP signaling pathway mediated by different TWF1 expression levels was a key hub in promoting the malignant transformation of LUAD cells. PKA could regulate the phosphorylation of Atg13 to strongly inhibit autophagy in Saccharomyces cerevisiae and this process occurred independently of the mTOR pathway. 43 In the pathogenesis of myocardial ischemiareperfusion injury, the adenosine A2a receptor (A2aR) exerted cardioprotective effects by inhibiting autophagy and apoptosis, which was regulated by the activation of cAMP-PKA signaling. 44 Taken together, the available evidence supports the cAMP signaling pathway as an autophagy-suppressive mechanism although the underlying mechanisms are not well understood. Therefore, the present study hypothesized that downregulation of TWF1 may inhibit autophagy via activation of the cAMP signaling pathway. As expected, the H89-2HCI-mediated inhibition of p-PKA and p-CREB1 significantly increased the levels of autophagy-related markers in the TWF1 downexpression groups, as confirmed by western blotting and LC3B immunofluorescence. The findings were indirectly supported by the report that the activation of the cAMP signaling pathway could inhibit autophagy-mediated degradation of HDAC8 in LUAD cells. 45 Notably, the in vitro findings in the present study were further supported by the subcutaneous xenograft model. In conclusion, TWF1 may promote the malignant progression of LUAD cells by inducing autophagy through inhibition of the cAMP/ PKA/CREB1 signaling pathway.
This part of the research still has several limitations. Firstly, further molecular biology experiments are needed to explore the underlying mechanism of the activation of the cAMP signaling pathway and anti-autophagy regulated by the downregulation of TWF1. Secondly, previous studies on the correlation between the TWF1 and autophagy and the cAMP signaling pathway were very limited, thus also limiting the exploration of the molecular mechanism in the present study to some extent. Therefore, the findings may provide a basis for further exploration of the downstream mechanisms of TWF1-induced LUAD progression.

| CONCLUSIONS
In summary, the results showed that low expression of TWF1 suppresses LUAD cells autophagy via regulation of the cAMP/PKA/CREB1 signaling pathway and interference with autophagy (Figure 7), which has some clinical implications. An an FDA-approved autophagy inhibitor, HCQ has shown a modest improvement in clinical response in several clinical trials targeting lung cancer. [46][47][48] F I G U R E 7 Proposed mechanistic scheme of TWF1 induces autophagy in LUAD.
In addition, two reviews have summarized several current clinical trials of anti-tumor drugs targeting the cAMP/ PKA signaling pathway. 38,49 Thus, understanding the role and mechanism of TWF1 in modulating autophagy and the cAMP/PKA/CREB1 signaling pathway could provide rationale and support for the treatment of LUAD in the future.

AUTHOR CONTRIBUTIONS
Hua Guo and Peng Chen designed the research studies; Yu Wang and Ran Zuo conducted experiments and performed the data analyses; Gengwei Huo, Zhiqiang Han, and Yuchao He helped perform analysis and manuscript preparation; Yi Luo, Liwei Chen, and Guangtao Li collated and counted clinical data; Jinfang Cui, Fuyi Zhu, and Ping Yue analyzed the clinical data; Dongqi Yuan, Yi Sun, and Zhaoyue Li contributed significantly to the smooth running of experiments; All authors read and critically revised the manuscript for intellectual content and approved the final manuscript.