Determining the Clinical Value and Critical Pathway of GTPBP4 in Lung

Hindawi BioMed Research International Volume 2020, Article ID 5171242, 13 pages https://doi.org/10.1155/2020/5171242

Research Article Determining the Clinical Value and Critical Pathway of GTPBP4 in Lung Adenocarcinoma Using a Bioinformatics Strategy: A Study Based on Datasets from The Cancer Genome Atlas

Zhiqian Zhang ,1 Juan Wang,2 Jiayan Mao,3 Fangqiong Li,2 Wei Chen ,3 and Wei Wang 2

1College of Medical Technology, Zhejiang Chinese Medical University, Hangzhou, Zhejiang 310053, China 2Department of Clinical Laboratory, Tongde Hospital of Zhejiang Province, Hangzhou, Zhejiang 310012, China 3Cancer Institute of Integrated Tradition Chinese and Western Medicine, Zhejiang Academy of Traditional Chinese Medicine, Tongde Hospital of Zhejiang Province, Hangzhou, Zhejiang 310012, China

Correspondence should be addressed to Wei Chen; [email protected] and Wei Wang; [email protected]

Received 4 May 2020; Revised 17 August 2020; Accepted 24 August 2020; Published 20 October 2020

Academic Editor: David A. McClellan

Copyright © 2020 Zhiqian Zhang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Lung cancer is the leading cause of cancer-related death worldwide, and the most common histologic subtype is lung adenocarcinoma (LUAD). Due to the significant mortality and morbidity rates among patients with LUAD, the identification of novel biomarkers to guide diagnosis, prognosis, and therapy is urgent. Guanosine triphosphate-binding protein 4 (GTPBP4) has been found to be associated with tumorigenesis in recent years, but the underlying molecular mechanism remains to be elucidated. In the present study, we demonstrate that GTPBP4 is significantly overexpressed in LUAD primary tumors. A total of 55 genes were identified as potential targets of GTPBP4. GO enrichment analysis identified the top 25 pathways among these target genes, among which, ribosome biogenesis was shown to be the most central. Each target gene demonstrated strong and complex interactions with other genes. Of the potential target genes, 12 abnormally expressed candidates were associated with survival probability and correlated with GTPBP4 expression. These findings suggest that GTPBP4 is associated with LUAD progression. Finally, we highlight the importance of the role of GTPBP4 in LUAD in vitro. GTPBP4 knockdown in LUAD cells inhibited proliferation and metastasis, promoted apoptosis, and enhanced sensitivity to TP. Overall, we conclude that GTPBP4 may be considered as a potential biomarker of LUAD.

1. Introduction ment, play an important role in the diagnosis, treatment, and prognosis of various cancer types [6, 7]. Adenocarcinoma is the most prevalent subtype of lung can- Guanosine triphosphate-binding protein 4 (GTPBP4), cer, comprising ~40% of all lung cancer cases [1, 2]. Despite also known as CRFG [8], NGB [9], and NOG1 [10], is a a wealth of research, lung adenocarcinoma (LUAD) remains GTPase involved in the synthesis of 60S ribosomal subunit as a highly aggressive and fatal disease, with an overall sur- and located on nuclear chromosome 10p15-14 [11]. Previous vival time of <5 years due to the difficulty in diagnosis [2– studies had shown that GTPBP4 can induce cell proliferation 4]. The discovery of novel specific molecular markers or tech- and enhance cell colony formation in some cancer types [12, nologies to diagnose LUAD is urgently required. Currently, 13]. A study of colorectal carcinoma (CRC) also uncovered computational biology is often combined with molecular that GTPBP4 was responsible for tumor metastasis [14]. biology and technology to explore the molecular mechanisms Meanwhile, patients with HCC (hepatocellular carcinoma) of disease and to identify clinically significant molecules [5]. with high levels of GTPBP4 expression tended to have a poor Biomarkers, usually molecules involved in cancer develop- prognosis [15]. The aforementioned research suggests that 2 BioMed Research International high expression of GTPBP4 is likely an important factor in 2.7. Correlation Analysis of GTPBP4 and the Selected Target the occurrence and development of tumors. However, there Genes. We used another TCGA analysis tool (http://gepia have been no prior reports regarding the expression or the .cancer-pku.cn/detail.php) to define the correlation between role of GTPBP4 in lung adenocarcinoma. GTPBP4 and the selected overlapping genes. In this study, we combined computational biology and experimental techniques to investigate the role of GTPBP4 2.8. Cell Lines and Cultures. Human LUAD cell lines (A549 in LUAD. The molecular function of GTPBP4 and its target and NCI-H1299) were purchased from the American Type genes were analyzed based on data from The Cancer Genome Culture Collection (ATCC; VA, USA) and cultured in RPMI Atlas (TCGA) database. We demonstrate that the high 1640 (Roswell Park Memorial Institute 1640) medium expression of GTPBP4 contributes to LUAD tumorigenesis (Gibco, Grand Island, NY) containing 1% penicillin/strepto- ff and might a ect the prognosis of patients. Additionally, mycin (Sigma, St Louis, MO) and 10% fetal bovine serum knockdown of GTPBP4 in A549 and H1299 cells inhibits ° (FBS, Gibco). All cells were maintained at 37 Cin5%CO2. proliferation and migration and improves cell apoptosis. Therefore, this study may provide a novel molecular target 2.9. Cell Transfection. GTPBP4 small interfering RNA for the treatment of LUAD. (siRNA) and negative control siRNA were synthesized obtained from GenePharma Co. (Shanghai, China). Briefly, 2. Materials and Methods cells at a density of 2:0×105 cells per well were transfected 2.1. GTPBP4 Expression in LUAD. The Cancer Genome Atlas with siRNA using Lipofectamine 2000 (Invitrogen, Carlsbad, (TCGA) is a high-throughput gene database containing data CA) in a serum-free medium in a 6-well plate. After 6 hours, regarding >30 types of human carcinoma [16]. We obtained the medium was replaced with fresh medium containing FBS. the GTPBP4 expression profiles for diverse cancer types and All subsequent experiments were performed at least 24 h with respect to gender, node metastasis status, ethnicity, and after transfection. The sequences of the GTPBP4 siRNA are ′ stage, based on a TCGA online analysis tool (http://ualcan as follows: GTPBP4-homo-245: sense: 5 -CCAUGAUAG .path.uab.edu/index.html). ACUUUCACAATT-3′ and antisense: 5′-UUGUGAAAG UCUAUCAUGGTT-3′; GTPBP4-homo-868: sense: 5′ 2.2. Prediction and Data Screening of GTPBP4 Target Genes. -GGGAGCAGCUAGAACUCUUTT-3′ and antisense: 5′ A total of 5 programs, including String (https://string-db.org/), -AAGAGUUCUAGCUGCUCCCTT-3′; GTPBP4-homo- BioGRID (https://thebiogrid.org/), BioPlex (https://bioplex.hms 1374: sense: 5′-GGCCAUAAUAUAGCUGAUUTT-3′ and .harvard.edu/), HPRD (http://www.hprd.org), and InBio_Map antisense: 5′-AAUCAGCUAUAUUAUGGCCTT-3′. (https://www.intomics.com/inbio/map/#home), were used to identify the target genes of GTPBP4. The target genes that overlapped in ≥3 of 5 programs were selected to improve the accu- 2.10. Cell Viability Assay. LUAD cells were seeded into 96- racy of the results. The integration of the genes was visualized well plates at a density of 5×103 cells per well and incubated by a Venn diagram (https://www.omicshare.com/tools/). overnight. Then, the medium was replaced with media containing different concentrations of Triptolide (TP; Sigma- 2.3. Functional Annotation of the Selected Target Genes in Aldrich) (0, 0.78125, 1.5625, 3.125, 6.25, 12.5, and 25 ng/mL) LUAD. Gene Ontology (GO) enrichment analysis was per- and incubated with the cells for 48 h. Cell viability was esti- formed to uncover the biological function of the overlapping mated using a Cell Counting Kit-8 assay (CCK8; Dojindo; target genes of GTPBP4 in LUAD. The online analysis tool Kumamoto, Japan), according to the manufacturer’s instruc- (https://www.omicshare.com/tools/) was applied to explore tion. A MAX II microplate reader (Dynex Technologies, the significance of the overlapping target genes of GTPBP4. Chantilly, VA) was used to measure the absorbance at 450 nm. The half maximal inhibitory concentration (IC50) 2.4. Protein-Protein Interaction (PPI) Network Construction. of TP was measured using the following equation: V%= The retrieval of interacting genes search tool (http:// 100/ð1+10½TPlogIC50Þ, where V% is the percentage viability metascape.org/gp/index.html#/main/step1) was used to dia- and [TP] is the TP concentration (ng/mL). gram the PPI network of the overlapping target genes. ′ 2.5. Identification of Differentially Expressed Genes. The 2.11. 5-Ethynyl-2 -deoxyuridine (EdU) Incorporation Assay. expression of the overlapping genes between LUAD primary DNA synthesis was quantified using a Click-iT™ EdU Imag- tumor and adjacent normal tissues was analyzed using a ing kit (Invitrogen; Carlsbad, CA, USA) according to the TCGA data online analysis tool (http://ualcan.path.uab.edu/ manufacturer’s instruction. Briefly, cells were seeded in 96- index.html). Results with statistical differences (P <0:05) well plates and 10 μM EdU was added to culture for 2 h. Next, were recorded as differentially expressed genes. cells were fixed with 4% paraformaldehyde for 15-30 min and permeabilized with 0.5% Triton X-100 for 20 min at room 2.6. Assessment of the Prognostic Value of the Overlapping temperature. After washing with PBS, 100 μL Click-iT reac- Genes of GTPBP4. The overlapping genes were searched on tion mixture was incubated with the cells for 30 min, TCGA data online analysis tool (http://ualcan.path.uab.edu/ followed by 100 μL Hoechst 33342 in PBS for 30 min. The index.html) to retrieve the associated survival curves results were visualized using a NanoZoomer 2.0-RS fluores- (P <0:05). cence microscope (Hamamatsu, Japan). BioMed Research International 3

Expression of GTPBP4 across TCGA cancers (with tumor and normal samples) 8

5 (TPM+1)

2 4 log 3

1 GBM KIRP KIRC LIHC CESC ESCA LUSC KICH SARC BLCA PCPG STAD BRCA PRAD READ UCEC HNSC PAAD LUAD SKCM CHOL THCA COAD THYM TCGA samples

Tumor Normal (a) Expression of GTPBP4 in LUAD Expression of GTPBP4 in LUAD based on sample types based on patient’s gender 100 100 ∗∗∗ ∗∗∗ 80 80

60 60

40 40 Transcript per million Transcript per million 20 20

0 0 Normal Primary tumor Normal Male Female (n = 59) (n = 515) (n = 59) (n = 238) (n = 276) TCGA samples TCGA samples (b) (c) Expression of GTPBP4 in LUAD Expression of GTPBP4 in LUAD 100 100 based on nodal metastasis status based on patient’s race

80 80

60 60

40 40

20 20 Transcript per million Transcript per million 0 0 Normal N0 N1 N2 N3 Normal Caucasian African- Asian n n n n n ( = 59) ( = 331) ( = 96) ( = 74) ( = 2) (n = 59) (n = 387) American (n = 8) TCGA samples (n = 51) TCGA samples (d) (e)

Figure 1: Continued. 4 BioMed Research International

Expression of GTPBP4 in LUAD based 100 on individual cancer stages

20 Transcript per million 0 Normal Stage1 Stage2 Stage3 Stage4 (n = 59) (n = 277) (n = 125) (n = 85) (n = 28) TCGA samples (f)

Figure 1: The expression proﬁle of GTPBP4 in LUAD clinical tissues. (a) GTPBP4 is overexpressed in the majority of human cancer types available via TCGA, including LUAD. (b) GTPBP4 expression is increased in primary tumor LUAD samples compared to normal samples. ∗∗∗ P <0:001. (c) The expression of GTPBP4 in males and females with LUAD. (d) The expression of GTPBP4 according to node metastasis status in patients with LUAD. N0: no regional lymph node metastasis; N1: metastases in 1-3 axillary lymph nodes; N2: metastases in 4-9 axillary lymph nodes; N3: metastases in ≥10 axillary lymph nodes. (e) The expression of GTPBP4 according to ethnicity among patients with LUAD. (f) The expression of GTPBP4 according to LUAD stage.

InBio_Map HPRD 30 min on ice. The soluble protein fractions were collected after centrifugation at 12000 × g for 20 min at 4°C, and pro- 275 tein concentration was quantified using a BCA protein assay 1 kit (Thermo Fisher; Rockford, IL, USA), according to the 17 manufacturer’s protocol. Equal amounts of protein were 5 2 0 0 denatured and separated by 10% SDS-PAGE and then trans- 0 7 ferred into polyvinylidene difluoride (PVDF) membranes 0 45 0 0 (Millipore; Billerica, MA, USA). PVDF membranes were blocked with 5% nonfat milk in TBST (Tris-buffered saline 3 0 0 0 (TBS) containing 0.1% Tween 20) at room temperature for 1 55 2 hours. Membranes were then incubated with anti- 20 4 GTPBP4 (13897-1-AP; Proteintech, Wuhan, China) and 3 9 0 1 anti-GAPDH (2118S; CST, Danvers, MA, USA), diluted 24 1 : 1000 in TBST overnight at 4°C. The membranes were then 0 8 0 30 washed 3 times and incubated with a horseradish peroxidase- BioGRID conjugated secondary antibody, diluted 1 : 2000 in TBST, at room temperature for 2 hours. Signals were visualized using BioPlex 34 ECL reagents (Thermo Scientific, Waltham, MA, USA), using GAPDH as a loading control. String 2.14. Scrape Motility Assay. The scrape motility assay was Figure 2: Venn diagram of the predicted target derived from 5 : 5 databases. used to evaluate cell migration. Cells (3 5×10 per well) were seeded in 6-well plates. Once the cells had formed confluent monolayers, a 200 μL sterile pipette tip was used to create a scratch in each well. The floating cells were removed, and 2.12. Cell Apoptosis Analysis. Preconditioned A549 and the anchorage-dependent cells were incubated in serum- H1299 cells were harvested and then washed twice with ice- free medium. Images were captured under an inverted light cold PBS. Then, Annexin V-FITC and PI dyes (BD Biosci- microscope (Olympus IX51, Olympus, Center Valley, PA, ences, Franklin Lakes, NJ, USA) were used to stain the cells, USA) at 0, 24, and 48 hours after scratching. according to the manufacturer’s description. Cells were finally analyzed via flow cytometry (BD Biosciences, Franklin Lakes, NJ, USA). 2.15. Statistical Analysis. Each experiment was performed independently ≥3 times. Results are presented as mean ± standard deviation (SD). All data were analyzed using Graph- 2.13. Western Blotting. Cells were washed twice with PBS and Pad Prism (version 8; GraphPad, San Diego, CA). Student’s t transferred to 100 μL lysis buffer (Beyotime Co, China) con- test was used to analyze differences between groups, and P < taining 100 mM phenylmethanesulfonyl fluoride (PMSF, 0:05 was considered to indicate a statistically significant Beyotime Co, China), in which they were incubated for difference. BioMed Research International 5

P Top 25 of GO term enrichment –log10 ( value) Ribosome biogenesis Ribonucleoprotein complex biogenesis rRNA metabolic process rRNA processing Ribosomal large subunit biogenesis ncRNA metabolic process 30 ncRNA processing RNA processing SRP-dependent cotranslational protein targeting to membrane Cotranslational protein targeting to membrane Protein targeting to ER Establishment of protein localization to endoplasmic reticulum Nuclear-transcribed mRNA catabolic process

GO term Nuclear-transcribed mRNA catabolic process, nonsense-mediated decay Protein localization to endoplasmic reticulum Translational initiation mRNA catabolic process RNA metabolic process 20 Nucleic acid metabolic process RNA catabolic process Viral transcription Protein targeting to membrane Viral gene expression Ribosomal large subunit assembly mRNA metabolic process

0.05 0.10 0.15 0.20 Rich factor Gene number 7 32

Figure 3: GO enrichment analysis of the identiﬁed overlapping target genes of GTPBP4 in LUAD.

3. Results 3.2. Prediction and Data Screening of GTPBP4 Target Genes. Groups of potential GTPBP4 target genes were selected from fi 3.1. GTPBP4 Expression in LUAD Tissues. To con rm the 5 public databases. A total of 119 genes were obtained from expression level of GTPBP4 in various human tumors, we String, 179 from BioGRID, 40 genes from BioPlex, 403 from used UALCAN, a TCGA data online analysis tool, to derive InBio_Map, and 33 from HPRD. This resulted in a total of fi the GTPBP4 expression pro les. GTPBP4 expression was 774 potential target genes. To improve prediction accuracy, increased in most human cancer types when compared with only genes obtained from ≥3 of 5 databases (overlapping ff adjacent normal tissues (Figure 1(a)), and this di erence was genes) were selected as potential targets of GTPBP4 for fur- fi signi cant in LUAD primary tumor tissues (Figure 1(b)). ther analysis, which was a total of 55 genes (Figure 2). GTPBP4 expression profiles in LUAD tissues were also compared for gender (male and female), node metastasis status (N0, N1, N2, and N3), ethnicity (Caucasian, African-Ameri- 3.3. Functional Analysis of the Overlapping GTPBP4 Target can, and Asian), and stage (S1, S2, S3, and S4). Statistical Genes in LUAD. The functional roles of 55 potential target analysis revealed that, in patients with LUAD, GTPBP4 genes in LUAD were analyzed in terms of biological pro- expression was higher in men than in women (Figure 1(c)). cesses (BP) by GO enrichment analysis. The top 25 enriched Excluding the impact of sample size, node metastasis status, pathways with significant differences were selected to con- ethnicity, and tumor stage did not result in the differential struct a bubble chart (Figure 3). The biological functions of expression of GTPBP4 (Figures 1(d)–1(f)). GTPBP4 target genes were mainly associated with RNA 6 BioMed Research International

RRS1 DDX24 MAK16 NOP53 DKC1 RPL27 GNL2 RPL4 NIFK DDX5 FBL NOP2 NOP58 MRTO4 WDR36 RPL14 DDX18 RPL11 ILF2 GNL3L RPL8 NMD3 CEBPZ RBM34 WDR12

RPF2 SDAD1 RPL7 WDR46

FN1 RPL18A BRIX1 RPL3 RPL10A DDX56 PWP1 ZC3H3 RBM28 RPL18 NOP56 NOC2L DDX55 RPL30 RPL37A PUM3 PES1 BOP1 RPL7A EIF6 NPM1 HIST1H1A NSA2 DDX47 NOL12PPAN

Figure 4: PPI network of the identified overlapping genes of GTPBP4. processing and metabolism, and the most significant path- to normal adjacent tissue except for DDX5 (Figure 5). In way was ribosome biogenesis (P <0:001). terms of LUAD prognosis, patients with downregulated DDX5 expression, or overexpression of NOP2, DDX18, 3.4. Protein-Protein Interaction (PPI) Enrichment Analysis of EIF6, BOP1, PES1, DDX47, RPF2, DDX56, MRTO4, RPL4, the Overlapping Target Genes. To determine the interaction and WDR46, had lower survival probabilities (Figure 6). Fur- among the proteins encoded by the overlapped genes in thermore, the correlation between the expression of GTPBP4 LUAD, Metascape software was used to construct a PPI net- and these 12 genes was calculated using GEPIA. In patients work. This was performed using 3 databases: BioGRID, with LUAD, NOP2, DDX18, EIF6, BOP1, PES1, DDX47, InWeb_IM, and OmniPath. As indicated in Figure 4, the RPF2, DDX56, MRTO4, RPL4, and WDR46 had a positive interactions among target genes are complicated, and each correlation with GTPBP4, while DDX5 had a negative corre- target gene has intricate associations with other genes. Most lation with GTPBP4 (Figure 7). of the targets belong to the PRL family, the DDX family, and the NOP family; these families all were reported to be 3.6. Knockdown of GTPBP4 Suppressed Cell Proliferation and essential for ribosome biogenesis and RNA metabolism, Accelerated Cell Apoptosis of LUAD Cells. To verify the results and this corresponds with the results in Section 3.3. of bioinformatics analysis, we analyzed the effect of downregulated GTPBP4 expression in LUAD cells in vitro. Western 3.5. Validation of the Overlapping Target Genes of GTPBP4 in blotting showed that siRNA targeting GTPBP4 (si- LUAD from TCGA Data. To further explore the role of GTPBP4-245, si-GTPBP4-868, and si-GTPBP4-1374) mark- GTPBP4 in LUAD, the expression of the overlapping target edly suppressed the expression of GTPBP4 in both A549 and genes and their association with prognosis were analyzed H1299 cells. The most striking downregulation of GTPBP4 using the UALCAN program. According to the query results, was achieved by si-GTPBP4-1374 (Figure 8(a)). CCK8 assays 12 genes with the most significant association between showed that RNA interference of GTPBP4 significantly expression and prognosis in LUAD were selected for further increased the sensibility of A549 and H1299 cells to TP com- analysis. Among the 12 genes (NOP2, DDX18, EIF6, BOP1, pared to control cells (Figure 8(b)). Based on the results of PES1, DDX47, RPF2, DDX56, MRTO4, RPL4, DDX5, and western blotting and CCK8 assays, si-GTPBP4-1374 was WDR46), all were upregulated in LUAD tissues compared selected for use in subsequent experiments. The EdU assays iMdRsac International Research BioMed ersnstmrsmls h xrsino I6 D1,MT4 O2 E1 P2 P4 D3,BP,DX7 n D5 is DDX56 LUAD. and in DDX47, decreased BOP1, was WDR36, DDX5 of RPL4, expression RPF2, the PES1, whereas NOP2, control, MRTO4, normal to DDX18, compared EIF6, LUAD of in expression increased The samples. tumor represents Figure Transcript per million Transcript per million Transcript per million Transcript per million 1000 1500 2000 2500 500 100 200 300 400 500 100 100 200 300 400 500 600 700 20 40 60 80 0 0 0 0 :Epeso fte1 ee nLA identifi LUAD in genes 12 the of Expression 5: Expression ofDDX5inLUAD Expression ofRPL4inLUAD Expression ofNOP2inLUAD Expression ofEIF6inLUAD Normal (n =59) (n =59) (n =59) (n =59) Normal Normal Normal based onsampletypes based onsampletypes based onsampletypes based onsampletypes TCGA samples TCGA samples TCGA samples TCGA samples Primary tumor Primary tumor Primary tumor Primary tumor (n =515) (n =515) (n =515) (n =515) ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗

Transcript per million Transcript per million Transcript per million Transcript per million 12.5 17.5 100 120 100 100 2.5 20 40 60 80 20 40 60 80 10 15 20 40 60 80 0 0 5 0 0 duigULA.Tebu oerpeet omlsmls n h e node red the and samples, normal represents node blue The UALCAN. using ed Expression ofWDR36inLUAD Expression ofDDX47inLUAD Expression ofDDX18inLUAD Expression ofPES1inLUAD Normal (n =59) (n =59) Normal (n =59) Normal Normal (n =59) based onsampletypes based onsampletypes based onsampletypes based onsampletypes TCGA samples TCGA samples TCGA samples TCGA samples Primary tumor Primary tumor Primary tumor Primary tumor (n =515) (n =515) (n =515) (n =515) ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗

Transcript per million Transcript per million Transcript per million Transcript per million 100 125 150 –10 25 50 75 10 20 30 40 10 20 30 40 50 60 70 10 20 30 40 50 60 70 0 0 0 0 Expression ofDDX56inLUAD Expression ofMRTO4inLUAD Expression ofBOP1inLUAD Expression ofRPF2inLUAD Normal Normal (n =59) (n =59) Normal (n =59) (n =59) Normal ∗∗∗ based onsampletypes based onsampletypes based onsampletypes based onsampletypes TCGA samples TCGA samples TCGA samples TCGA samples P <0 : 001 Primary tumor Primary tumor Primary tumor Primary tumor . (n =515) (n =515) (n =515) (n =515) ∗∗∗ ∗∗∗ ∗∗∗ ∗∗∗ 7 8 BioMed Research International

Efect of EIF6 expression level Efect of DDX18 expression level Efect of MRTO4 expression level on LUAD patient survival on LUAD patient survival on LUAD patient survival 1.00 1.00 1.00

0.75 0.75 0.75

0.50 0.50 0.50

0.25 0.25 0.25 P Survival probability P Survival probability P = 0.0094 = 0.042 Survival probability = 0.024

0.00 0.00 0.00 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Time in days Time in days Time in days

Expression level Expression level Expression level High expression (n = 125) High expression (n = 126) High expression (n = 126) Low/medium expression (n = 377) Low/medium expression (n = 376) Low/medium expression (n = 376)

Efect of NOP2 expression level Efect of PES1 expression level Efect of RPF2 expression level on LUAD patient survival on LUAD patient survival on LUAD patient survival 1.00 1.00 1.00

0.75 0.75 0.75

0.50 0.50 0.50

0.25 0.25 0.25 Survival probability Survival probability Survival probability P = 0.006 P = 0.015 P = 0.012

0.00 0.00 0.00 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Time in days Time in days Time in days

Efect of RPL4 expression level Efect of WDR36 expression level Efect of BOP1 expression level on LUAD patient survival on LUAD patient survival on LUAD patient survival 1.00 1.00 1.00

0.75 0.75 0.75

0.50 0.50 0.50

0.25 0.25 P 0.25 Survival probability Survival probability Survival probability P = 0.0043 = 0.015 P = 0.026

0.00 0.00 0.00 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Time in days Time in days Time in days

Expression level Expression level Expression level High expression (n = 126) High expression (n = 126) High expression (n = 126) Low/medium expression (n = 376) Low/medium expression (n = 376) Low/medium expression (n = 376)

Efect of DDX5 expression level Efect of DDX47 expression level Efect of DDX56 expression level on LUAD patient survival on LUAD patient survival on LUAD patient survival 1.00 1.00 1.00

0.75 0.75 0.75

0.50 0.50 0.50

0.25 P = 0.026 0.25 P = 0.016 0.25 P < 0.0001 Survival probability Survival probability Survival probability

0.00 0.00 0.00 0 2000 4000 6000 0 2000 4000 6000 0 2000 4000 6000 Time in days Time in days Time in days

Expression level Expression level Expression level High expression (n = 126) High expression (n = 125) High expression (n = 125) Low/medium expression (n = 376) Low/medium expression (n = 377) Low/medium expression (n = 377)

Figure 6: The prognostic signiﬁcance of the 12 selected target genes for LUAD. Low expression of DDX5 and high expression of EIF6, DDX18, MRTO4, NOP2, PES1, RPF2, RPL4, WDR36, BOP1, DDX47, and DDX56 were associated with poor prognosis of LUAD patients. BioMed Research International 9

P value = 0 7 P P value = 0 value = 0 6 R = 0.33 R R = 0.48 8 6 = 0.48 5 5 6 4 4 3 4 3 (EIF6 TPM) 2 (DDX18 TPM) (MRTO4 TPM) 2 2 2

log 2 log 2 log 1 1 000 01234567 01234567 01234567

log2 (GTPBP4 TPM) log2 (GTPBP4 TPM) log2 (GTPBP4 TPM) 6 7 P value = 0 P value = 0 P value = 0 R = 0.49 R = 0.43 R = 0.5 6 6 5 5 4 4 4 3 3 (PES1 TPM) (RPF2 TPM) (NOP2 TPM) 2 2 2 2 2 log

2 log log 1 1 0 0 0 01234567 01234567 01234567

log2 (GTPBP4 TPM) log2 (GTPBP4 TPM) log2 (GTPBP4 TPM)

P value = 1.5e−14 P value = 0 P value = 0 12 4 R = 0.26 R = 0.34 R = 0.46 10 6 3 8 4 6 2 (RPL4 TPM) (BOP1 TPM) 2 2 (WDR36 TPM)

4 2 log

log 2

log 1 2

000 01234567 01234567 01234567

log2 (GTPBP4 TPM) log2 (GTPBP4 TPM) log2 (GTPBP4 TPM) P 8 10 value = 0.063 7 P value = 0 P value = 0 R = −0.065 R = 0.39 R = 0.32 6 8 6 5 6 4 4 4 3 (DDX5 TPM) (DDX47 TPM) (DDX56 TPM) 2 2 2 2 2 log 2 log log 1 0 0 0 01234567 01234567 01234567

log2 (GTPBP4 TPM) log2 (GTPBP4 TPM) log2 (GTPBP4 TPM)

Figure 7: Validation of the 12 selected target genes using the GEPIA server. EIF6, DDX18, MRTO4, NOP2, PES1, RPF2, RPL4, WDR36, BOP1, DDX47, and DDX56 all positively correlated with GTPBP4 in LUAD, whereas DDX5 was negatively correlated with GTPBP4. revealed that proliferation of A549 and H1299 cells was sup- 3.7. GTPBP4 Facilitates the Migration of LUAD Cells. To pressed by si-GTPBP4 (Figure 8(c)). Similarly, downregula- investigate whether GTPBP4 may contribute to LUAD tion of GTPBP4 expression induced apoptosis of both A549 metastasis, siRNA was used to downregulate GTPBP4 and H1299 cells (Figure 8(d)). expression in A549 and H1299 cells. Downregulation of rlfrto ae fLA el eeaaye sn h d sa iha IC an with assay EdU the using analyzed di were under cells determined LUAD of was rates cells Proliferation H1299 and A549 of 10 TB425 iGPP-6,ads-TB417 a con was si-GTPBP4-1374 and si-GTPBP4-868, GTPBP4-245, Figure ppoi a vlae yAnxnVstaining; V Annexin by evaluated was apoptosis NCI-H1299 A549 PI-A PI-A si-GTPBP4 si-NC Control GTPBP4 10 10 10 10 GAPDH 10 10 10 10 2 3 4 5 2 3 4 5 :GPP euae UDcl rlfrto n ppoi.()Kokono TB4i 59adH29clsuigs-C si- si-NC, using cells H1299 and A549 in GTPBP4 of Knockdown (a) apoptosis. and proliferation cell LUAD regulates GTPBP4 8: 96.2% 0.2% 95.6% 0.1%0.1 95.6 EdU 10 10 % 2 2 % 10 10 Control Control FITC-A FITC-A 3 3 3Q4 Q3 Q2 Q1 Q3Q Q 3Q4 Q3 Q2 Q1 A549 3 1 si-NC 10 10 Q4Q Q2Q si-GTPBP4-245 DAPI A549 4 2 4 4 (a) si-GTPBP4-868 2.1% 1.5% 0.8 3.5 0.8% 3.5% 10 10 si-GTPBP4-1374 % % 5 5

Control NCI-H1299 PI-A PI-A 10 10 10 10 10 10 10 10

Merge si-NC 2 3 4 5 2 3 4 5 94.7%94. 0.3% 93.0% .%4.3% 0.1% 0.1 si-GTPBP4-245 10 10 50 �m 7% % 2

2 si-GTPBP4-868 %

Annexin V si-GTPBP4-1374 10 10 en±SD ± mean FITC-A FITC-A si-NC 3 3 3Q4 Q3 Q2 Q1 Q3Q Q 3Q4 Q3 Q2 Q1

1 si-GTPBP4 si-NC Control ﬀ Relative cell viability rn ocnrtoso P(5 25 .5 .2,152,0715 n gm) (c) ng/mL). 0 and 0.78125, 1.5625, 3.125, 6.25, 12.5, (25, TP of concentrations erent 10 Q4Q QQ2 10 4 2 4 4 (% of control) EdU 100 1.3% 20 40 60 80 0.9% 0.9 5.5% 4 0 .3 10 10 ( % 01 025 20 15 10 5 0 % n 5 5 ﬁ =3 mdb etr ltig(et 59cls ih:H29cls.()Viability (b) cells). H1299 right: cells; A549 (left: blotting western by rmed si-GTPBP4-245 si-NC Control

), PI-A PI-API-A 10 10 10 10 10 10 10 10 Triptolide (ng/mL) NCI-H1299 ∗∗∗ 2 3 4 5 (d) 2 3 4 5 (c) 1.6% 81.2% 85.5% 8 DAPI 0% 0% 5. 10 10 P 5% A549 2 2 <0 si-GTPBP4 10 : 10 001 FITC-A FITC-A 3 3 3Q4 Q3 Q2 Q1 Q4 QQ3 Q3 Q2 Q1 50 3 . ocnrto fTP. of concentration Merge Q4Q Q2Q 10 10 4 2 4 4 si-GTPBP4-1374 si-GTPBP4-868 50 �m 10.8% 12.9% 1 6.5% 1.6% 2.9 1.61 10 10 (b) % % 5 5

PI Relative cell viability EdU-positive cell ratio (%) (% of control) 20 40 60 80 100 0 20 40 60 80 0

Apoptosis (%) 25 20 15 10 5 0 iMdRsac International Research BioMed ∗ 10 15 20 25 P 0 5 si-GTPBP4 si-NC Control <0 A549 Triptolide (ng/mL) : 05 ∗∗ si-GTPBP4 si-NC Control NCI-H1299 A549 and ∗∗∗ ∗∗ NCI-H1299 P <0 NCI-H1299 ∗ : 01 d Cell (d) . ∗∗∗ BioMed Research International 11

A549 NCI-H1299

Control si-NC si-GTPBP4 Control si-NC si-GTPBP4

0 h 0 h

24 h 24 h

48 h 48 h

200 �m 200 �m

A549 NCI-H1299 80 50 40 60 30 40 20

20 10

0 0 Wound healing rate (%) Wound healing rate (%) 24 48 24 48 –10 –20 Time (h) Time (h)

Control Control si-NC si-NC si-GTPBP4 si-GTPBP4 (a) (b)

Figure 9: GTPBP4 promotes migration of LUAD cells. Wound healing was monitored for 48 h in A549 and H1299 cell monolayers; mean ± SD (n =3).

GTPBP4 expression in A549 cells partially inhibited the [12, 13], it has also been described as a suppressor gene in migration of LUAD cells 24 h posttransfection, and this inhi- rare cases, such as in neurofibromatosis 2 (NF2) [9], suggest- bition was magnified at 48 h after transfection. In H1299 ing that the role of GTPBP4 depends on the specific type of cells, downregulated GTPBP4 expression also repressed cancer. metastasis 24 h and 48 h posttransfection compared with In the present study, we have demonstrated that the the control and si-NC groups. There was no difference in expression of GTPBP4 is higher in tumor tissues than in metastasis between the control and si-NC groups. These adjacent normal control. Similarly, the immunohistochemis- results indicated that GTPBP4 facilitated the migration of try (IHC) database of HPA (https://www.proteinatlas.org) LUAD cells (Figure 9). showed that the protein expression of GTPBP4 in LUAD tissues was increased compared to normal control (Figure S1). 4. Discussion In addition, to further investigate the function of GTPBP4, its potential downstream target genes were identified. We GTPBP4 is a molecular switch that is of great importance for show that GTPBP4 may be a novel potential biomarker to the biogenesis of the 60S ribosomal subunit and signal trans- guide LUAD diagnosis, prognosis, and treatment. mission due to its GTPase activity [8, 17, 18]. It switched Suppression of GTPBP4 expression in A549 and H1299 between an active state and inactive state when bound with cells inhibited cell proliferation and migration and GTP or GDP [19]. At a time when targeted therapy is popu- increased the rate of apoptosis. Based on the anti-tumor lar in the field of cancer research, GTBBP4 has been identi- activity of TP and its toxicity to LUAD cells [20], our study fied as a potential biomarker for various cancer types. indicated that RNA interference of GTPBP4 in A549 and However, recent studies have indicated that the role of H1299 cells significantly increased sensibility to TP. GTPBP4 in malignant tumor types is double sided. Although Some studies have reported that GTPBP4 plays a critical GTPBP4 has been most often found to act as an oncogene role in the progression of tumors. For example, aberrantly 12 BioMed Research International expressed GTPBP4 was found to be significantly associated Data Availability with low survival probability in patients with HCC and breast fi cancer [15, 21]. Yu et al. concluded that GTPBP4 was respon- The data used to support the ndings of this study are avail- sible for tumor metastasis in CRC [14], and Li et al. suggested able from the corresponding author upon request. that GTPBP4 promotes gastric cancer progression [12]. The analysis of its potential target genes in the present study sug- Conflicts of Interest gested that GTPBP4 may be a predictor of survival in LUAD fl cases. The overall survival analysis graph indicates that high The authors declare that there is no con ict of interest expression of GTPBP4 in LUAD is associated with poor regarding the publication of this paper. prognosis; however, this was not statistically significant (Figure S2). This also implies that GTPBP4 may predict Authors’ Contributions prognosis of patients with LUAD. As shown in Figure 3, ribosome biogenesis was predicted Zhiqian Zhang, Juan Wang, and Jiayan Mao contributed to be the crucial biological process involving GTPBP4. equally to this work. Importantly, a series of studies have indicated that dysregu- lated ribosome biogenesis is essential for the tumorigenesis Acknowledgments of most spontaneous cancer types and that ribosome biogenesis is closely related to tumor suppressor P53 in cell prolifer- This study was supported by the National Science Founda- ation and apoptosis [22–24]. The majority of the selected tion (No. 81774026). target genes in the present study were also observed to participate in ribosome synthesis. EIF6 (eukaryotic initiation factor Supplementary Materials 6) has been shown to be essential for nucleolar biogenesis of Figure S1: IHC of GTPBP4 in LUAD sample and normal 60S ribosomes and maximal protein synthesis downstream of fi growth factor stimulation [25]. The assembly of 40S and 60S lung sample from HPA, ×100 magni cation. Figure S2: sur- ribosomal subunits is regulated by DEAD-box RNA helicase vival curve analysis based on GTPBP4 expression in patients with LUAD; the high expression of GTPBP4 was associated 18 (DDX18) [26]. Nucleolar protein 2 (NOP2/NSUN1), P : which can inhibit HIV-1 transcription and promote viral with poor prognosis of LUAD patients ( =007). latency [27], is also required for nucleolar maturation and (Supplementary Materials) ribosome biogenesis in mammals [28]. Ribosomal protein L4 (RPL4) is known to affect tumorigenesis and metastasis References of HCC [29] and to participate in the assembly of pre-60S [1] R. Remark, C. Becker, J. E. Gomez et al., “The non-small cell in the nucleus [30]. BOP1 (Block of Proliferation 1) is lung cancer immune contexture. A major determinant of responsible for modulating pre-rRNA processing of 28S tumor characteristics and patient outcome,” American Journal and 5.8S rRNAs [31, 32]. Additionally, DEAD-box RNA heli- of Respiratory and Critical Care Medicine, vol. 191, no. 4, case 56 (DDX56) is involved in the assembly of the 60S large pp. 377–390, 2015. ribosomal subunit and has been associated with lymphatic [2] T. V. Denisenko, I. N. Budkevich, and B. Zhivotovsky, “Cell invasion and distant metastasis in CRC [33]. Pescadillo ribo- death-based treatment of lung adenocarcinoma,” Cell Death somal biogenesis factor 1 (PES1), which has been reported as & Disease, vol. 9, no. 2, p. 117, 2018. an independent poor prognostic factor in pancreatic cancer [3] X. Lei, L. du, P. Zhang et al., “Knockdown GTSE1 enhances patients [34], is essential for 60S ribosomal subunit matura- radiosensitivity in non–small-cell lung cancer through DNA tion and pre-rRNA processing [35]. Therefore, we speculated damage repair pathway,” Journal of Cellular and Molecular that the regulation of ribosome biogenesis by GTPBP4 plays Medicine, vol. 24, no. 9, pp. 5162–5167, 2020. an essential role in various types of cancer. Regarding pro- [4] M. Früh, D. De Ruysscher, S. Popat, L. Crinò, S. Peters, and moted progression of gastric cancer by GTPBP4 regulation E. Felip, “Small-cell lung cancer (SCLC): ESMO Clinical Prac- ff tice Guidelines for diagnosis, treatment and follow-up,” of P53 activity [12], we infer that aberrant GTPBP4 a ects – LUAD cell proliferation, apoptosis, and migration by dis- Annals of Oncology, vol. 24, pp. vi99 vi105, 2013. “ fi turbing the balance between ribosome biogenesis and P53 [5] Y. Shen, J. Liu, L. Zhang et al., Identi cation of potential bio- activity. Thus, GTPBP4 is likely to participate in LUAD by markers and survival analysis for head and neck squamous cell carcinoma using bioinformatics strategy: a study based on regulating ribosomal biogenesis. TCGA and GEO datasets,” BioMed Research International, vol. 2019, Article ID 7376034, 14 pages, 2019. [6] F. Skoulidis, M. E. Goldberg, D. M. Greenawalt et al., 5. Conclusions “STK11/LKB1Mutations and PD-1 inhibitor resistance inKRAS-Mutant lung adenocarcinoma,” Cancer Discovery, In conclusion, our findings suggest that aberrantly high vol. 8, no. 7, pp. 822–835, 2018. expression of GTPBP4 contributes to the tumorigenesis of [7] O. Calvayrac, A. Pradines, E. Pons, J. Mazieres, and N. Guibert, LUAD and that it may be associated with prognosis of “Molecular biomarkers for lung adenocarcinoma,” The Euro- patients with LUAD. GTPBP4 may have potential as a novel pean Respiratory Journal, vol. 49, no. 4, p. 1601734, 2017. therapeutic target, a diagnostic biomarker, or a survival pre- [8] N. J. Laping, B. A. Olson, and Y. Zhu, “Identification of a novel dictor for LUAD. nuclear guanosine triphosphate-binding protein differentially BioMed Research International 13

expressed in renal disease,” Journal of the American Society of [24] D. Awad, M. Prattes, L. Kofler et al., “Inhibiting eukaryotic Nephrology, vol. 12, no. 5, pp. 883–890, 2001. ribosome biogenesis,” BMC Biology, vol. 17, no. 1, p. 46, 2019. [9] H. Lee, D. Kim, H. C. Dan et al., “Identification and character- [25] E. Pesce, A. Miluzio, L. Turcano et al., “Discovery and prelim- ization of putative tumor suppressor NGB, a GTP-binding inary characterization of translational modulators that impair protein that interacts with the neurofibromatosis 2 protein,” the binding of eIF6 to 60S ribosomal subunits,” Cells, vol. 9, Molecular and Cellular Biology, vol. 27, no. 6, pp. 2103–2119, no. 1, p. 172, 2020. 2007. [26] H. Zhang, Z. Wu, J. Y. Lu et al., “DEAD-box helicase 18 coun- [10] S. Lee, M. Senthil-Kumar, M. Kang et al., “The small GTPase, teracts PRC2 to safeguard ribosomal DNA in pluripotency nucleolar GTP-binding protein 1 (NOG1), has a novel role in regulation,” Cell Reports, vol. 30, no. 1, pp. 81–97.e7, 2020. ” fi plant innate immunity, Scienti c Reports, vol. 7, no. 1, p. 9260, [27] W. Kong, A. Biswas, D. Zhou et al., “Nucleolar protein 2017. NOP2/NSUN1 suppresses HIV-1 transcription and promotes [11] B. C. Jensen, Q. Wang, C. T. Kifer, and M. Parsons, “The viral latency by competing with Tat for TAR binding and NOG1 GTP-binding protein is required for biogenesis of the methylation,” PLOS Pathogens, vol. 16, no. 3, article 60 S ribosomal subunit,” The Journal of Biological Chemistry, e1008430, 2020. – vol. 278, no. 34, pp. 32204 32211, 2003. [28] H. Wang, L. Wang, Z. Wang et al., “The nucleolar protein [12] L. Li, X. Pang, Z. Zhu et al., “GTPBP4 promotes gastric cancer NOP2 is required for nucleolar maturation and ribosome bio- progression via regulating P53 activity,” Cellular Physiology genesis during preimplantation development in mammals,” and Biochemistry, vol. 45, no. 2, pp. 667–676, 2018. PLoS Pathogens, vol. 34, no. 2, pp. 2715–2729, 2020. [13] M. J. van de Vijver, Y. D. He, L. J. van 't Veer et al., “A gene- [29] X. Yang, L. Sun, L. Wang, B. Yao, H. Mo, and W. Yang, expression signature as a predictor of survival in breast can- “LncRNA SNHG7 accelerates the proliferation, migration cer,” The New England Journal of Medicine, vol. 347, no. 25, and invasion of hepatocellular carcinoma cells via regulating pp. 1999–2009, 2002. miR-122-5p and RPL4,” Biomedicine & Pharmacotherapy, [14] H. Yu, S. Jin, N. Zhang, and Q. Xu, “Up-regulation of GTPBP4 vol. 118, article 109386, 2019. in colorectal carcinoma is responsible for tumor metastasis,” [30] B. Pillet, J. J. Garcia-Gomez, P. Pausch et al., “The dedicated Biochemical and Biophysical Research Communications, chaperone Acl4 escorts ribosomal protein Rpl4 to its nuclear vol. 480, no. 1, pp. 48–54, 2016. pre-60S assembly site,” PLoS Genetics, vol. 11, no. 10, article [15] W. B. Liu, W. D. Jia, J. L. Ma et al., “Knockdown of GTPBP4 e1005565, 2015. inhibits cell growth and survival in human hepatocellular car- [31] M. Rohrmoser, M. Holzel, T. Grimm et al., “Interdependence cinoma and its prognostic significance,” Oncotarget, vol. 8, of Pes1, Bop1, and WDR12 controls nucleolar localization no. 55, pp. 93984–93997, 2017. and assembly of the PeBoW complex required for maturation ” [16] S. Guo, J. Yang, M. Wu, and G. Xiao, “Clinical value screening, of the 60S ribosomal subunit, Molecular and Cellular Biology, – prognostic significance and key pathway identification of miR- vol. 27, no. 10, pp. 3682 3694, 2007. 204-5p in endometrial carcinoma: a study based on The Can- [32] C. S. Ahn, H. K. Cho, D. H. Lee, H. J. Sim, S. G. Kim, and H. S. cer Genome Atlas (TCGA), and bioinformatics analysis,” Pai, “Functional characterization of the ribosome biogenesis Pathology - Research and Practice, vol. 215, no. 5, pp. 1003– factors PES, BOP1, and WDR12 (PeBoW), and mechanisms 1011, 2019. of defective cell growth and proliferation caused by PeBoW fi ” [17] S. Wu, B. Tutuncuoglu, K. Yan et al., “Diverse roles of assem- de ciency in Arabidopsis, Journal of Experimental Botany, – bly factors revealed by structures of late nuclear pre-60S ribo- vol. 67, no. 17, pp. 5217 5232, 2016. somes,” Nature, vol. 534, no. 7605, pp. 133–137, 2016. [33] Y. Kouyama, T. Masuda, A. Fujii et al., “Oncogenic splicing fi [18] H. Shi, N. M. Chapman, J. Wen et al., “Amino acids license abnormalities induced by DEAD-box helicase 56 ampli cation ” kinase mTORC1 activity and Treg cell function via small G in colorectal cancer, Cancer Science, vol. 110, no. 10, – proteins Rag and Rheb,” Immunity, vol. 51, no. 6, pp. 1012– pp. 3132 3144, 2019. 1027.e7, 2019. [34] X. Jin, R. Fang, P. Fan et al., “PES1 promotes BET inhibitors [19] A. K. Gillingham, J. Bertram, F. Begum, and S. Munro, “In vivo resistance and cells proliferation through increasing c-Myc ” identification of GTPase interactors by mitochondrial relocali- expression in pancreatic cancer, Journal of Experimental & zation and proximity biotinylation,” eLife, vol. 8, 2019. Clinical Cancer Research, vol. 38, no. 1, p. 463, 2019. “ [20] L. Chen, Q. Liu, Z. Huang et al., “Tripchlorolide induces cell [35] W. Xie, Q. Feng, Y. Su et al., Transcriptional regulation of ” death in lung cancer cells by autophagy,” International Journal PES1 expression by c-Jun in colon cancer, PLoS One, vol. 7, of Oncology, vol. 40, no. 4, pp. 1066–1070, 2012. no. 7, article e42253, 2012. [21] A. Lunardi, G. Di Minin, P. Provero et al., “A genome-scale protein interaction profile of Drosophila p53 uncovers addi- tional nodes of the human p53 network,” Proceedings of the National Academy of Sciences of the United States of America, vol. 107, no. 14, pp. 6322–6327, 2010. [22] J. Pelletier, G. Thomas, and S. Volarevic, “Ribosome biogenesis in cancer: new players and therapeutic avenues,” Nature Reviews Cancer, vol. 18, no. 1, pp. 51–63, 2018. [23] M. Derenzini, L. Montanaro, and D. Trere, “Ribosome biogenesis and cancer,” Acta Histochemica, vol. 119, no. 3, pp. 190– 197, 2017.