We mainly using the following databases to analysis the expression, prognosis, clinical information and Immune infiltration of SNX20 in cancers. The detail information of databases used in this study are as follows: Oncomine database (https://www.oncomine.org/) (Rhodes et al., 2004), TIMER (http://timer.cistrome.org/) (Li et al., 2017), UALCAN (http://ualcan.path.uab.edu/) (Chandrashekar et al., 2017), Kaplan-Meier plotter (http://kmplot.com/) (Nagy et al., 2021), TISIDB database (http://cis.hku.hk/TISIDB/) (Ru et al., 2019), the Human Disease Methylation Database (http://bio-bigdata.hrbmu.edu.cn/) (Lv et al., 2012), the PrognoScan database (http://dna00.bio.kyutech.ac) (Mizuno et al., 2009), The Linked Omics database (http://www.linkedomics.org/) (Vasaikar et al., 2018). GEPIA database (http://gepia.cancer-pku.cn/) (Tang et al., 2017) and cbioportal (http://www.cbioportal.org/) (Cerami et al., 2012).

We employ the starbase (www.starbase.sysu.edu.cn) (Rhodes et al., 2004), Targetscan (http://www.targetscan.org/) (Li et al., 2017), miRDB (http://mirdb.org) (Chandrashekar et al., 2017) and miRbase (http://microrna.sanger.ac.uk/) (Nagy et al., 2021) to predict the potential miRNAs that able to binds with miRNAs. We also using the starbase (Rhodes et al., 2004) to analysis the expression and prognosis of miRNAs and lncRNAs. The starbase (www.starbase.sysu.edu.cn) (Nagy et al., 2021) and lncRNASNP (http://bioinfo.life.hust.edu.cn/lncRNASNP/) (Lv et al., 2012) was apply to analysis the upstream lncRNAs that bindings with the miRNAs. Furthermore, we using the starbase to analysis the correlation between the LncRNA/miRNA/mRNA.

We employed the STRING (www.string-db.org) (Mering, 2003) and GeneMANIA (http://www.genemania.org) (Warde-Farley et al., 2010) perform protein-protein and gene-gene interaction network analysis of SNX20.

We employed the lncLocator (http://www.csbio.sjtu.edu.cn/bioinf/lncLocator2/) (Cao et al., 2018) and CPC2 (http://cpc2.cbi.pku.edu.cn) analysis the Cell localization and coding potential of lncRNA.

The cDNA for SNX20 was constructed employ pCDH-CMV-MCS-EF1-Puro vector. The NSCLC related cells mainly purchased from The Kunming Institute of Zoology (KIZ) of Chinese Academy of Sciences (CAS). BEAS-2B and NSCLC related cells lines were cultured using the RPMI1640 medium, this medium contain the 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. The cDNA for SNX20 primer sequences are list follows: SNX20-PCR-F: ATG​GCA​AGT​CCA​GAG​CAC​CCT​G, SNX20-PCR-R: TCA​GGG​TGT​GGC​GTC​AGG​AGC​CGG​AGC​CA.

The qRT-PCR assay was performed as documented (Jiang et al., 2018). For Real-time RT-PCR assay, indicated cells were lysed by RNAiso Plus (Takara Bio, Beijing, China, Cat# 108-95-2). Total RNA was extracted according to the manufacturer’s protocol, and then reverse transcribed using RT reagent Kit (Takara Bio, Beijing, China, Cat# RR047A; TIANGEN Biotech, Beijing, China, Cat# KR211-02). Real-time PCR was performed by FastStart Universal SYBR Green Master Mix (Roche, Cat# 04194194001; TIANGEN Biotech, Beijing, China, Cat# FP411-02) using an Applied Biosystems 7,500 machine. The primers and antibodies used in this study are shown in xx. The primer sequences are list follows SNX20-F: ACC​TGA​CGG​GCA​CTT​AGA​CA, SNX20-R: AGA​GCA​GTT​TGA​CGT​GCT​TCC; β-actin-F: CTTCGCGGGCGACGAT, β-actin-R: CCA​TAG​GAA​TCC​TTC​TGA​CC. The expression quantification was obtained with the 2−ΔΔCt method.

Cell proliferation, colony formation, tumor sphere formation assay was performed as documented (62). Briefly, for cell proliferation assay, indicated cells were plated into 12-well plates at a density of 1.5 × 104, the cell numbers were subsequently counted each day using an automatic cell analyzer countstar (Shanghai Ruiyu Biotech Co., China, IC 1000). For colony formation assay, indicated cells were seeded in 6-well plate with 600 cells per well supplemented with 2 ml cell culture medium, and the cell culture medium was changed every 3 days for 2∼3 weeks. Indicated cells were fixed with 4% PFA and stained with 0.5% crystal violet.

Cell migration assays was performed as documented (62). Briefly, to produce a wound, the monolayer cells in 6-well plate were scraped in a straight line with pipette tips. Plate was then washed with warm PBS to remove detached cells. Photographs of the scratch were taken at indicated time points using Nikon inverted microscope (Ti-S). Gap width was calculated with GraphPad Prism software. For trans-well assay, 2.5−3×104 cells in 100 μL serum free medium were plated in an 8.0 μm, 24-well plate chamber insert (Corning Life Sciences, catalog no. 3422), with medium containing 10% FBS at the bottom of the insert. Cells were incubated for 24 h, and then fixed with 4% paraformaldehyde for 20 min. After washing, cells were stained with 0.5% crystal violet blue. The positively stained cells were examined under the microscope.

Briefly, putative binding sites for miR-301a-3p on the 3′-UTR of SNX20 was predicted by starbase dataBase (http://starbase.sysu.edu.cn/). Wild-type and mutant SNX20 (mut- SNX20 or mut- SNX20) fragments were constructed and inserted downstream of the luciferase reporter gene in the reporter plasmid pGL3 plasmid (Promega). HEK-293T cells (2 × 104 cells/well) were seeded in a 24-well plate and co-transfected with 3′-UTR SNX20 construct and miR-301a-3p mimics or miR Ctrl using Lipofectamine 3,000. Both firefly and Renilla luciferase expressions were measured post-transfection using the Dual Luciferase Kit (Promega) according to the manufacturer’s instructions.

The Western Blotting and Immunohistochemistry staining assay was performed as documented (Griffiths-Jones, 2006). Briefly, Cell lysates were collected, perform the Western blot, primary antibody overnight incubation and second antibody incubation. Finally, develop using instrument.The detail information of antibodies employ in our study are as follows: SNX20 antibody (SNX20, FNab08087, Rabbit 1:1,000,) and β-actin (Catalog number: 66009-1-Ig, 1:20,000, Proteinch, Shanghai, China).

The significance of the data between two assays groups was decided by Student’s t-test, p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***), was considered significantly.

In order to examine the mRNA of SNX20 expression pattern in multifarious cancer, we employed the TIMER tools to analysis the expression of SNX20, the result shown that SNX20 was low expression in LUAD, LUSC, and PAAD, higher expression was observed in BRCA, CHOL, ESCA, GBM, HNSC, KIRC, and KIRP (Figure 1A). To further verify the results, we using the combine the TCGA and GTE databases to figure out the SNX20 expression. As is show in Figure 1B, the SNX20 was significantly up-regulation in COAD, GBM, KIRC, KIRP, LAML, PAAD, SKCM, SARC, and TGCT cancer than match healthy tissue. Besides, we found that the SNX20 was down regulated in NSCLC cells lines observe in CCLE network tools (Figure 1C). Above all, our findings indicated that the SNX20 may plays different roles in the progression of different cancers.

We employ the GEPIA tools to examine the relationship between the expression of SNX20 and the human cancers pathological stage. Interestingly, we find that the expression of SNX20 was markedly positive with the pathological stage of KIRC, SARC and negative with the pathological stage of OV and THYM (Figures 2A–D). These results suggested that SNX20 plays different roles in different human cancers.

The prognostic value of SNX20 expression in human cancers was analyzed by several databases. In GEPIA, we found that lower SNX20 expression was associated with poorer overall survival (OS) in CESC, LUAD and SARC, the patients with higher SNX20 expression had poor OS in LGG and UVM (Figures 3A–E). Additionally, high expression of SNX20 has a better DFS observed in SKCM, LIHC and UCEC (Figures 3F–H). Lower expression of SNX20 was related to poor DSS in CESC, HNSC, LGG, LUAD, SKCM, UCEC and UVM (Supplementary Figure S1A), and linkage to poor PFS in CESC, HNSC, KIRP and UCEC (Supplementary Figure S1B). The above results proved that SNX20 expression closely related to the prognosis of various cancer types.

Considering the significance of SNX20R in cancers, next, we want to exploration the correlation between SNX20 expression and clinical features in LUAD. First, we found that the lower expression was observed in the three GEO datasets (Figures 4A–C). Next, we find that the RNA of SNX20 was significantly lower in LUAD by perform the UALCAN tools analysis ((Figure 4D). In addition, we also find the expression of SNX20 was decreased with the elevation of stage nodal metastasis and tumor stage (Figures 4E,F). Surprisingly, we find higher expression of SNX20 has the better OS, PFS and PPS (Figures 4G–I). Finally, we adopt the GEO dataset to verify above results, the studies shown that elevated the SNX20 expression display the better OS and RFS in NSCLC (Figures 4J,K). ROC curve analysis of SNX20 showed an AUC value of 0.884 in TCGA LUAD patients (Figure 4L). These results indicate that SNX20 has the potential to act as a detection index for the diagnosis of lung cancer with high sensitivity and specificity.

For explore the gene mutation information about the SNX20, we employ the cBioportal tools preform comprehensive analysis regard to the SNX20. The result shown that the mutation rate of SNX20 reached 1.8% in NSCLC (Figures 5A,B), the results also display the mutation of SNX20 in different NSCLC molecular Subtypes (Figure 5C), Next, we also examine the mutation type and base mutation in NSCLC, we found that Missense substitution and base G > A reached the highest mutation rate in NSCLC (Figures 5D,E). Overall, these results emphasize the gene mutation of SNX20 may be contribute to the SNX20 low expression in NSCLC.

As a one of the crucial epigenetic modification, DNA Methylation plays an significant roles in regulation gene expression. In order to explore the molecular mechanism of the SNX20 aberrantly up-regulated in LUAD, we analyzed the promoter methylation levels of SNX20 in LUAD. We found that there are many methylation sites in the promoter region of SNX20, and the differential methylation regions were indicated in the heatmaps (Figure 6A). Importantly, by using the shiny methylation analysis resource tool (SMART) analysis (Tay et al., 2014) we uncovered that the methylation of SNX20 was significantly higher in LUAD cancerous tissues compared to that in normal tissues (Figure 6B). Consistently, we found that the methylation levels on the specific methylation site (cg06207201) within SNX20 promote region negatively correlated with its expression in LUAD (Figure 6C). Furthermore, we showed that the elevated methylation levels on cg06207201 site correlates with worse OS in the TCGA-LUAD cohorts, using the methSurv dataset (Figure 6D). Additionally, we also found that another two methylation sites (cg01144086 and cg08330349) also negatively with the expression of SNX20 in LUAD (Figure 6E). Above all, there results suggested the DNA methylation plays crucial roles in the modulates the expression of SNX20 in LUAD.

To further study the function of SNX20, we employed the Linkedomics to perform a correlation analysis of SNX20. Two heatmaps were constructed to illustrate the genes whose expression was most positively and negatively correlated with that of SNX20 (Figures 7A,B). GO annotation revealed that these genes participate in various immune response, including T cell activation, interferon γ production, adaptive immune response, leukocyte proliferation, regulation of defense response to virus by virus, myeloid dendritic cell activation, interleukin-4 production, leukocyte activation involved in inflammatory response, lymphocyte activation involved in immune response, response to chemokine, immune response regulating signaling pathway and regulation of leukocyte activation (Figure 7C). KEGG pathway analysis showed the enrichment in intestinal immune network for IgA production, primary immunodeficiency, hematopoietic cell lineage, T cell receptor signaling pathway, cell adhesion molecules, Th1 and Th2 cell differentiation, Th17 cell differentiation, Natural killer cell medicated cytotoxicity and Fc epsilon RI signaling pathway (Figure 7D). Furthermore, we also employed the GeneMania and STRING databases construction the gene and protein interaction networks of SNX20 in LUAD. The results indicated that the twenty frequently altered genes was correlated with SNX20, including SELPLG, ENOSF1 and NCF (Figure xx), the significantly correlated with the SNX20 protein mainly including ZNRF2, METTL6, FKBP14, HECA and SELPLG (Figures 7E,F).

To further explore the molecular mechanisms affected by SNX20 in Lung adenocarcinoma, we perform the GSEA enriched analysis and find SNX20 mainly participate in the immune related biology processes. GO enriched results indicated that SNX20 were enriched mainly in regulation of adaptive immune response, positive regulation of immune effector process, CD4 positive β T cell activation and B cell differentiation (Figure 8A). Similarly, The KEGG enriched results shown that SNX20 were enriched mainly involve in chemokine signaling pathway, JAK STAT signaling pathway, T cell receptor signaling pathway and Toll like receptor signaling pathway (Figure 8B). These results strongly indicated that SNX20 was mainly involved in regulation the immune response of lung adenocarcinoma.

To investigate whether some miRNAs regulates the SNX20 expression by miRNA sponge manner, we employed the starbase (Li et al., 2014), Targetscan (Agarwal, 2015), miRDB (Wong and Wang, 2015) and miRbase (Griffiths-Jones, 2006) to predict the potential miRNAs that able to binds with miR-301a-3p. The results find that has-miR-34c-5p, has-miR-301a-3p, has-miR-338-5p, and has-miR-3614-5p may be bind with SNX20. Next, we employed the Starbase and Kmplot databases analysis the expression and prognosis of SNX20 in Lung cancer (Figures 9A,B). According to competing endogenous RNAs hypothesis, the miRNA expression would be negative with the mRNA expression (Tay et al., 2014). Among there miRNAs, only has-miR-301a-3p expression was negatively correlated with SNX20 in LUAD patients (Figure 9C), we also analysis the base pair relationship between the miR-301a-3p and SNX20 (Figure 9D). By perform comprehensive analysis the expression, prognosis and expression correlation, we confirmation the miR-301a-3p may be as aN miRNA sponge for SNX20 (Figure 9E). Furthermore, we adopt the NSCLC GEO dataset to verify the expression of miR-301a-3p in lung cancer, the analysis results shown that miR-301a-3p expression lower in the whole blood and tissue in human lung cancer (Figures 9F,G). To investigate the effects of miR-301a-3p on the expression of target gene, we conducted overexpression analysis. Overexpressing miR-301a-3p significantly reduced both mRNA and protein expressions of SNX20 in A549 cell (Figures 9H–J). The luciferase assay showed that transfection of miR-30a-5p mimics significantly reduced the relative luciferase activity of SNX20-3UTR-WT-treated lung cancer cells, but did not affect that of SNX20-3UTR-MUT-treated lung cancer cells (Figure 9K). All these results confirmed the tumor suppressor roles of miRNA-30a-5p in the lung cancer progression.

To explore the upstream lncRNAs that binding with miR-301a-3p, we employed the starbase (Nagy et al., 2021) and lncRNASNP (Gong et al., 2015) analysis the upstream lncRNAs. We find 4 possible lncRNA, including the AC01086.4, XIST, SNX20AR and SLC26A4-AS1. We first examine the expression and prognosis of lncRNAs by employed the starbase and Kmplot (Figures 10A,B). According to competing endogenous RNAs hypothesis, the miRNA expression would be negative with the lncRNA expression, and lncRNA should positive correlation with the mRNA expression (Vasaikar et al., 2018). Owing to the mRNA was decreased expression in LUAD and miRNA was highly expression in LUAD, the lncRNA should be low expression in LUAD. By perform comprehensive analysis the expression, prognosis and expression correlation, we confirmation the SNX20AR may be as an lncRNA sponge for miR-301a-3p. Next, the correlation analysis indicated that the SNX20AR expression not only negatively correlated with miR-301a-3p, but also positively correlated with SNX20 in LUAD (Figures 10C,D).We also analysised the base pair relationship between the miR-301a-3p and SNX20AR (Figure 10F). Finally, we find that the SNX20AR expression was decreased with the cancer stage elevated, and adopt the NSCLC GEO dataset to verify the expression of SNX20AR in lung cancer, the analysis results shown that SNX20AR expression lower in the in human lung cancer (Figures 10G,H). In order to explored the molecular characteristics of SNX20AR, we perform the related analysis by employed the lncLocator (Cao et al., 2018) and Coding Potential Calculator (CPC) (Kong et al., 2007) find that the SNX20AR not only mainly in the cytoplasm, but also no possess protein coding potential (Figures 10I,J). Thus, the data show that SNX20AR may be as a sponge for miR-33a-5p in LUAD.

To explore the significances of SNX20 in the tumor microenvironment of LUAD. We employed the TIMER to analysis the correlation of SNX20 level with immune infiltration in Lung adenocarcinoma. We find that SNX20 CNV was significantly affect the immune infiltration levels of the immune cells, including the B cell, CD8⁺T cells, CD4⁺T cells, Macrophage, Neutrophil and Dendritic cell (Figure 11A). In addition, we also show that the expression of SNX20 were negatively correlated with tumor purity (r = −0.509, p = 5.96e-34) and positively associated with the immune infiltration of B cells (r = 0.579, p = 1.39e-44), CD8⁺ T cell (r = 0.451, p = 9.14e-26), CD4⁺ T cell (r = 0.685, p = 2.66e-28), Macrophage (r = 0.557, p = 6.29e-41), Neutrophil (r = 0.801, p = 2.07e-110) and Dendritic cell (r = 0.328, p = 5.46E-10) in LUAD (Figure 11B). Additionally, we also employed the cox proportional hazard model examine the prognostic value of SNX20 expression and tumor infiltration immune cells in LUAD. The result suggested that B cells (p = 0) and CD4⁺ T cells (p = 0.014) were significantly associated with clinical prognosis in LUAD (Table 1).

Undoubtedly, studies has been demonstrated that immune checkpoints play an crucial roles in the development of cancer immunotherapy (Haanen and Robert, 2015). We employed the TIMER database to exploration the correlation between the SNX20 expression and immune check point related gene. The results suggested that SNX20 was markedly positive with the expression of immune check point related gene, including the CTLA4 (r = 0.685, p = 1.03e-72), CD274 (r = 0.646, p = 4.81e-62), HAVCR2 (r = 0.804, p = 8.43e-118), LAG3 (r = 0.519, p = 6.62e-37), PDCD1 (r = 0.668, p = 6.25e-68), PDCD1LG2 (r = 0.754, p = 1.19e-95), SIGLEC15 (r = 0.371, p = 3.18e-18) and TIGIT (r = 0.758, p = 2.38e-97) (Figure 11C). Finally, we also employed TISIDB database analysis the relationship between SNX20 level and 28 tumors immune infiltrating cell subtypes and the immune regulator (Table 2). The results suggested that SNX20 was positively with the 28 tumor immune infiltrating cell, immune regulator and MHC molecular (Supplementary Figure S2). These findings indicate that SNX20 plays significant role in cancer immune regulation of LUAD.

We estimated the correlation between SNX20 expression and different immune cell gene marker in LUAD by employed the TIMER database. Our results showed that the SNX20 expression was strongly correlation with the different immune markers. For instances, SNX20 expression was strongly correlation with CD8⁺ T markers, CD8A (r = 0.607), CD8B (r = 0.495). The correlation with between the SNX20 expression and immune cells markers as is show in the Table 3. These findings indicated that SNX20 was participate in regulate the tumor immune infiltration in lung adenocarcinoma.

Due to the SNX20 expression can affect the immune infiltration of immune cells. We further explored the expression of SNX20 and the different immune cells infiltration whether influence the prognosis of lung cancer patients. The analysis results suggested that patients with the high expression of SNX20 and enriched the B cells, CD4⁺ cells, CD8⁺ cells, Eosinophils cells, Macrophages cells, Mesenchymal stem cells, Type 1 helper cells and Type 2 helper cells will display a better prognosis. While, The highly expression SNX20 and decreased the B cells, CD4⁺ cells, CD8⁺ cells, Eosinophils cells, Macrophages cells, Mesenchymal stem cells, Type 1 helper cells and Type 2 helper cells will display a poor prognosis (Figures 12A–D). Together, these results suggested that SNX20 expression and different immune cells infiltration would be affect the prognosis of LUAD.

In order to explored the function of SNX20 in LUAD progression. We first examine the SNX20 mRNA expression in NSCLC cells lines. The result show that the mRNA and protein levels of SNX20 were decreased in NSCLC cells, especially in A549 and H1299 cells (Figures 13A,B). Considering the SNX20 was low expression in NSCLC cells, we construction stable overexpression SNX20 cells and using the qRT-PCR assay detection the over-expression efficiency (Figure 13C). Furthermore, we perform the gain of function to examine elevated the SNX20 expression whether affect the proliferation and migration ability of NSCLC cells. The growth curve and clone information experimental result shown that elevated SNX20 was significantly suppressed the cell growth ability of NSCLC cells (Figures 13D–F). Similarly, decreased the migration ability of NSCLC cells was observed in over-expression of SNX20 NSCLC cells (Figures 13G–K). The above findings indicated that SNX20 plays tumor suppressor role in the LUAD progression.

Above results suggested that SNX20 may plays oncogene roles in the cancer progression, so we next explored the correlation between SNX20 expression and different drug sensitivity in different cancer cell lines from the GDSC and CTRP database. The result indicated that SNX20 expression was negatively correlated with drug sensitivity of I-BET-762, KIN001-260, CAL-101, PIK-93, PHA-793887, BIX02189, TPCA-1, NG-25, QL-XI-92, KIN001-244, TG101348, TL-1-85, BHG712 and NPK76-II-72-1in GDSC database (r < −0.36, Figure 14A). In CTRP database, we observed SNX20 expression was negatively correlated with the drug sensitivity of teniposide, BRD-K30748066, SR-II-138A, Narciclasine, GSK461364, CR-1-31B, LY-2183240, LRRK2-IN-1, SNX-2112, piperlongumine, ciclopirox, PX-12 and triazolothiadiazine (r < −0.35, Figure 14B). In summary, these results demonstrate that SNX20 was significantly related to diverse drug sensitivity in the different cancer cell lines.