Identification of hub genes specific to pulmonary metastasis in osteosarcoma through integrated bioinformatics analysis

Abstract

BACKGROUND:

Pulmonary metastasis is the most frequent cause of death in osteosarcoma (OS) patients. Recently, several bioinformatics studies specific to pulmonary metastatic osteosarcoma (PMOS) have been applied to identify genetic alterations. However, the interpretation and reliability of the results obtained were limited for the independent database analysis.

OBJECTIVE:

The expression profiles and key pathways specific to PMOS remain to be comprehensively explored. Therefore, in our study, three original datasets of GEO database were selected.

METHODS:

Initially, three microarray datasets (GSE14359, GSE14827, and GSE85537) were downloaded from the GEO database. Differentially expressed genes (DEGs) between PMOS and nonmetastatic osteosarcoma (NMOS) were identified and mined using DAVID. Subsequently, GO and KEGG pathway analyses were carried out for DEGs. Corresponding PPI network of DEGs was constructed based on the data collected from STRING datasets. The network was visualized with Cytoscape software, and ten hub genes were selected from the network. Finally, survival analysis of these hub genes also used the TARGET database.

RESULTS:

In total, 569 upregulated and 1238 downregulated genes were filtered as DEGs between PMOS and NMOS. Based on the GO analysis result, these DEGs were significantly enriched in the anatomical structure development, extracellular matrix, biological adhesion, and cell adhesion terms. Based on the KEGG pathway analysis result, these DEGs were mainly enriched in the pathways in cancer, PI3K-Akt signaling, MAPK signaling, focal adhesion, cytokine-cytokine receptor interaction, and IL-17 signaling. Hub genes (ANXA1 and CXCL12) were significantly associated with overall survival time in OS patient.

CONCLUSION:

Our results may provide new insight into pulmonary metastasis of OS. However, experimental studies remain necessary to elucidate the biological function and mechanism underlying PMOS.

Keywords

Osteosarcoma pulmonary metastasis differentially expressed genes pathways bioinformatics ANXA1

1. Introduction

Osteosarcoma (OS) is most common type of malignant bone cancer. The pulmonary tissue is the most common site for metastatic spread of the disease, accounting for more than 75% of metastases. Pulmonary metastasis is the major cause of OS leading deaths [1, 2, 3, 4]. Hence, novel therapeutic targets for pulmonary metastasis in osteosarcoma (PMOS) are urgently needed.

Many reports demonstrated that abnormal genes contribute to the pathogenesis [5, 6, 7, 8] and metastasis of OS [9, 10, 11, 12, 13]. Recently, several bioinformatics studies specific to pulmonary metastasis have been applied to identify genetic alterations in PMOS. However, DEGs in these reports are mostly based on tissue-cell mixed samples or independent database analysis, limiting the interpretation and reliability of the results obtained. The expression profiles and key pathways specific to PMOS remain to be comprehensively explored. Therefore, in our study, three original datasets of GEO database were selected according to strict screening criteria: microarray datasets that contain both PMOS and nonmetastatic osteosarcoma (NMOS) in human tissue samples. We mined and analyzed data from these databases, anticipating the identification of gene signatures specific to PMOS and creating a foundation for further studies.

2. Materials and methods

2.1 Data sources

Microarray data were downloaded from the publicly available GEO database. The screening indicators were as follows: data that contain both PMOS and nonmetastatic osteosarcoma (NMOS), human tissue samples. Datasets that did not meet these indicators were excluded. After careful view, three datasets were identified for further analysis: GSE14359 [14], GSE14827 [15, 16], and GSE85537 [17] (Table 1). Raw data and probe annotation files were downloaded for further analysis.

Table 1
Summary of microarray datasets

Dataset	Platform	Total samples	Non-metastasis samples	Pulmonary metastasis samples	PMID	Country
GSE14359	GPL96-affymetrix	18	10	8	21166698	Germany
GSE14827	GPL570-affymetrix	27	18	9	20159990, 24448647	Japan
GSE85537	GPL570-affymetrix	6	3	3	28944885	China

GEO, Gene expression Omnibus; PMID, PubMed unique identifier.

2.2 Screening DEGs

The GEO2R online analysis tool (https://www.ncbi.nlm.nih.gov/geo/geo2r/) was used to identify genes between PMOS and NMOS samples. Genes (only mRNA were screened out) that met the cutoff criteria, $p<$ 0.05 and $|$ logFC $|$ $>$ 2.0, were considered as DEGs. Statistical analysis and Venn diagram webtool were performed.

2.3 GO and KEGG analyses

The GO analysis include the classified type of BP, MF, and CC. GO and KEGG pathway analyses of DEGs used DAVID tools (https://david.ncifcrf.gov/). $p<$ 0.01 and gene count $\geqslant$ 20 were considered statistically significant.

2.4 Construction of PPI network

The DEGs that were previously identified were mapped to the STRING database (http://string-db.org/) to evaluate possible PPI relationships. PPI pairs were extracted for an interaction score of 0.9. PPI network was visualized using Cytoscape (3.8.1).

2.5 Identification of hub genes

CytoHubba was used to calculate the node degree for each protein. The top 10 genes were identified as hub genes.

2.6 Survival analysis

The correlation between key genes and survival time was assessed with the survival function in the R package using clinical information from the TARGET database. Patients in this series matrix were divided into high and low expression groups based on the expression level of mRNA. In overall survival analysis of each hub gene, GraphPad Prism software (v.8, https://www.graphpad.com/scientific-software/prism/) was used. Log-rank $p$ -value was calculated, and $p<$ 0.05 was considered statistically significant.

Figure 1.

Volcano plot of DEGs. Note: Yellow, blue, and red color represent the relatively down, non-different and up expression of genes, respectively. X: $|$ logFC $|$ $>$ 2; Y: $p<$ 0.05.

Figure 2.

GO analysis of DEGs. Note: (A) Go annotation of DEGs in three parts: BP, MF, and CC. The horizontal axis represents the name of the GO terms and the vertical axis represents the number of genes. (B) The top 20 of enrichment terms of DEGs. The horizontal axis represents the rich factor value and the vertical axis represents the name of the GO terms. The size of the dot represents the number of genes involved in the terms. The color changes from red to green represent changes in significance from high to low.

Figure 3.

KEGG pathway analysis of DEGs. Note: (A) KEGG pathway annotation of DEGs in six parts: Cellular Processes, Human Diseases, Environmental Information Processing, Organismal Systems, Metabolism, and Genetic Information Processing. The horizontal axis represents the number of genes and the vertical axis represents the name of the pathways. (B) The top 20 of enrichment pathways of DEGs. The horizontal axis represents the rich factor value and the vertical axis represents the name of the pathways. The size of the dot represents the number of genes involved in the pathways. The color changes from red to green represent changes in significance from high to low.

3. Results

3.1 Identification of DEGs

After a careful screening, a total of 340 DEGs were identified from GSE14359, including 268 upregulated and 72 downregulated genes; 67 DEGs were identified from GSE14827, including 5 upregulated and 62 downregulated genes; 1426 DEGs were identified from GSE85537, including 304 upregulated and 1122 downregulated genes. All DEGs were identified by comparing PMOS samples with NMOS samples. 26 DEGs was found to be differentially expressed in at least two profiles. Unfortunately, zero DEGs was found to be differentially expressed in all three profiles. Therefore, we harvested 1807 genes after summing up the three databases and removing the duplicates. Subsequently, volcano plot analysis of these DEGs was performed (Fig. 1).

3.2 Functional enrichment analysis

GO and KEGG pathway analyses were carried out to explore the functions of these 1807 DEGs. GO analysis indicated that DEGs were enriched in numerous functional terms (Fig. 2A). BP terms included cellular process, biological regulation, regulation of BP, metabolic processes, response to stimulus, multicellular organismal processes, developmental processes, signaling, localization, and positive regulation of BP. MF terms included binding, catalytic activity, molecular function regulator, transcription regulator activity, and transporter activity. CC terms included cell, organelle, membranes, extracellular regions, and protein-containing complexes. A bubble chart for these terms was organized into the following: anatomical structure development, developmental processes, multicellular organism development, multicellular organismal processes, extracellular matrix, system development, anatomical structure morphogenesis, collagen-containing extracellular matrix, animal organ development, cellular developmental processes, cell development, biological adhesion, cell adhesion, cell differentiation, regulation of developmental processes, extracellular regions, extracellular matrix organization, plasma membrane regions, regulation of multicellular organismal processes, and extracellular structure organization as the top 20 enrichment GO terms (Fig. 2B).

KEGG analysis indicated that DEGs were enriched in numerous pathways (Fig. 3A). The bubble chart for these pathways was organized into the following: cancer, PI3K-Akt signaling, MAPK signaling, cytokine-cytokine receptor interaction, IL-17 signaling, focal adhesion, AGE-RAGE signaling in diabetic complications, ECM-receptor interaction, Rap1 signaling, proteoglycans in cancer, protein digestion and absorption, ovarian steroidogenesis, relaxin signaling, calcium signaling, TNF signaling, Chagas disease (American trypanosomiasis), cAMP signaling, platelet activation, transcriptional misregulation in cancers, and glutamatergic synapse (Fig. 3B).

3.3 PPI network construction

PPI network was constructed using the STRING database to examine interactions among DEGs. A visualized network was created by Cytoscape with confidence $>$ 0.9 and hiding the disconnected nodes. The final network consisted of 704 nodes and 6060 edges (Fig. 4).

Figure 4.

PPI network. Note: Pink nodes represent upregulated genes, blue nodes represent downregulated genes.

3.4 Hub gene identification

The identified top 10 genes, which were evaluated by degree of connectivity in the PPI network, were as follows: complement C3 (C3, degree $=$ 62), G protein subunit gamma 11 (GNG11, degree $=$ 61), pro-platelet basic protein (PPBP, degree $=$ 56), annexin A1 (ANXA1, degree $=$ 49), G protein-coupled receptor 17 (GPR17, degree $=$ 46), C-X-C motif chemokine ligand 8 (CXCL8, degree $=$ 45), C-X-C motif chemokine ligand 12 (CXCL12, degree $=$ 42), adenylate cyclase 1 (ADCY1, degree $=$ 42), epidermal growth factor receptor (EGFR, degree $=$ 42), and angiotensin II receptor type 2 (AGTR2, degree $=$ 41) (Table 2). A subnetwork of these top 10 genes was constructed to explore connections (Fig. 5). C3, PPBP, EGFR, and AGTR2 were upregulated in PMOS, and GNG11, ANXA1, GPR17, CXCL8, CXCL12, and ADCY1 were downregulated.

Table 2
Top ten hub genes with higher degree of connectivity

Ensemble ID	Gene symbol	Gene description	Degree
ENSG00000125730	C3	Complement C3	62
ENSG00000127920	GNG11	G protein subunit gamma 11	61
ENSG00000163736	PPBP	Pro-platelet basic protein	56
ENSG00000135046	ANXA1	Annexin A1	49
ENSG00000144230	GPR17	G protein-coupled receptor 17	46
ENSG00000169429	CXCL8	C-X-C motif chemokine ligand 8	45
ENSG00000107562	CXCL12	C-X-C motif chemokine ligand 12	42
ENSG00000164742	ADCY1	Adenylate cyclase 1	42
ENSG00000146648	EGFR	Epidermal growth factor receptor	42
ENSG00000180772	AGTR2	Angiotensin II receptor type 2	41

Figure 5.

Top 10 hub genes. Note: 10 identified hub genes: C3, GNG11, PPBP, ANXA1, GPR17, CXCL8, CXCL12, ADCY1, EGFR, AGTR2. Red (high degree value), yellow (low degree value).

3.5 Survival analysis

Modulation of ANXA1 ( $p=$ 0.0142) and CXCL12 ( $p=$ 0.0109) expression were significantly correlated with the overall survival of OS patients (Fig. 6).

Figure 6.

Overall survival analysis of the hub genes. Note: Gene changes of ANXA1 (A) and CXCL12 (B) were significantly correlated with the overall survival time of OS patients ( $P<$ 0.05).

4. Discussion

Previous studies are mostly based on a single dataset or tissue-cell mixed datasets, limiting the interpretation of the results in PMOS [18, 19, 20, 21]. In the present study, we strictly integrated three datasets according to the screening criteria to identify key genes correlated with PMOS. By comparing the three mRNA profiles downloaded from the GEO database, 569 upregulated and 1238 downregulated DEGs were screened among 20 PMOS tissue samples and 31 NMOS tissue samples. These DEGs were associated with many GO terms significantly enriched in various KEGG pathways. GO and KEGG analyses indicated that DEGs were closely associated with tumor progression and chemoresistance [22, 23, 24, 25, 26]. Subsequently, PPI network was constructed to investigate the interrelationship among the DEGs. The top 10 hub genes were ranked based on the degree of connectivity. C3, PPBP, EGFR, and AGTR2 were upregulated in PMOS, whereas GNG11, ANXA1, GPR17, CXCL8, CXCL12, and ADCY1 were downregulated. Finally, a survival curve was generated to assess relationships between the expression level of hub genes and prognosis of OS patients. Low circulating levels of ANXA1 and CXCL12 were linked to significantly worse survival.

ANXA1, a member of the annexin superfamily, is a calcium-dependent phospholipid-binding protein. The role of ANXA1 in tumor development and metastasis has been documented in many cancers [27, 28, 29, 30, 31, 32, 33], but its role and underlying mechanism in OS are poorly understood. We show for the first time a clinically positive correlation between ANXA1 and overall survival rate.

CXCL12 (also known as SDF-1), a member of the CXC subfamily of chemokines, is secreted by stromal cells from a variety of tissues [34]. The involvement of CXCL12 is implicated in tumor progression of a variety of cancers and immune diseases [35, 36, 37], but its role in the metastasis of OS is ambiguous. An interaction between CXCL12 and its two receptors, CXCR4 and CXCR7, drives the metastasis of OS cells to the lung [38, 39, 40, 41]. Interestingly, CXCL12 is weakly, or even negatively, expressed in lung metastases from OS [42]. CXCL12 positively correlated with better longtime survival and lower prevalence of metastasis [43]. High expression of CXCL12 in OS tissues predicted longer survival, which is consistent with our current results. Osteoblasts and OS cells express the specific chemokine axis (CXCR4, CXCR7, and CXCL12) might prevent tumor cells from entering the circulation [44, 45]. Recently, the levels of CXCL4, CXCL6, and CXCL12 in plasma in OS patients were significantly higher compared with that in the controls. However, the higher circulating levels of CXCL4 and CXCL6, but not CXCL12, were associated with a poorer outcome for OS patients [46]. Certainly, the CXCL12-CXCR4/CXCR7 axis is not the sole regulator of systemic spread but could be crucial for metastatic cancer development. Moreover, PI3K pathways are common important downstream pathways regulated by ANXA1 [47] and the CXCL12-CXCR4/CXCR7 axis [48].

5. Conclusion

The present study identified genes that may be useful diagnostic biomarkers and treatment goals for PMOS patients. The results will support studies of mechanisms and effective therapies for PMOS. However, due to limitations in the present study, additional work is required to elucidate molecular mechanisms involved in this challenging disease. Necessarily, further research is required to verify the mechanisms of ANXA1 and CXCL12 in PMOS.

Author contributions

Data collection: Yinan Chai; Data analysis: Lihan Xu, Rui He; Manuscript writing: Liangjun Zhong; Conception and design: Yuying Wang.

Data availability

The data used to perform the analyses described herein are publicly available from the GEO (www.ncbi.nlm.nih.gov/geo) and TARGET (https://ocg.cancer.gov/programs/target) databases.

Ethical considerations

There was no need to obtain ethical approval from the hospital since all data in the manuscript was based on bioinformatics, and GEO and TARGET have obtained informed consent from all participants.

Funding

This work was supported by the Natural Science Foundation of Zhejiang Province (Grant number: LY20H060003).

Footnotes

Conflict of interest

The authors declare that there are no competing interests associated with the manuscript.

References

Huang

Hou

Yang

. The metastasectomy and timing of pulmonary metastases on the outcome of osteosarcoma patients. Clin Med Oncol. 2009 Sep 14; 3: 99-105.

Meazza

Scanagatta

. Metastatic osteosarcoma: A challenging multidisciplinary treatment. Expert Rev Anticancer Ther. 2016 May; 16(5): 543-56.

Morrow

Bayles

Funnell

APW

Miller

Saiakhova

Lizardo

Bartels

Kapteijn

Hung

Mendoza

Dhillon

Chee

Myers

Allen

Gambarotti

Righi

DiFeo

Rubin

Huang

Meltzer

Helman

Picci

Versteeg

Stamatoyannopoulos

Khanna

Scacheri

. Positively selected enhancer elements endow osteosarcoma cells with metastatic competence. Nat Med. 2018 Feb; 24(2): 176-185.

Ahmed

Zamzam

Kamel

Ahmed

Salama

Zaki

Kamal

Elshafiey

. Effect of timing of pulmonary metastasis occurrence on the outcome of metastasectomy in osteosarcoma patients. J Pediatr Surg. 2019 Apr; 54(4): 775-779.

Zhao

Zhang

Tan

Wang

Liu

Song

Wang

. Gene function analysis in osteosarcoma based on microarray gene expression profiling. Int J Clin Exp Med. 2015 Jul 15; 8(7): 10401-10.

Peng

Yang

Wei

Yuan

Liu

Yin

Wang

Zhang

Kong

. Investigation of crucial genes and microRNAs in conventional osteosarcoma using gene expression profiling analysis. Mol Med Rep. 2017 Nov; 16(5): 7617-7624.

Chen

Wang

Yang

Zhang

Song

. Potential regulatory effects of miR-182-3p in osteosarcoma via targeting EBF2. Biomed Res Int. 2019 Mar 12; 2019: 4897905.

Liu

Xie

Wang

Lei

. Identification of potential crucial genes and key pathways in osteosarcoma. Hereditas. 2020 Jul 14; 157(1): 29.

Wang

. Identification of biomarkers for metastatic osteosarcoma based on DNA microarray data. Neoplasma. 2015; 62(3): 365-71.

10.

Sun

Shen

Yin

Wang

Cai

. Bioinformatics analysis of differentially expressed pathways related to the metastatic characteristics of osteosarcoma. Int J Mol Med. 2016 Aug; 38(2): 466-74.

11.

. A support vector machine classifier for the prediction of osteosarcoma metastasis with high accuracy. Int J Mol Med. 2017 Nov; 40(5): 1357-1364.

12.

Fan

Wang

Zhang

. Identification of critical genes associated with human osteosarcoma metastasis based on integrated gene expression profiling. Mol Med Rep. 2019 Aug; 20(2): 915-930.

13.

Tang

Cao

Peng

Zhao

Zeng

Gao

Guo

. Overexpression of Notch3 is associated with metastasis and poor prognosis in osteosarcoma patients. Cancer Manag Res. 2019 Jan 8; 11: 547-559.

14.

Fritsche-Guenther

Noske

Ungethüm

Kuban

Schlag

Tunn

Karle

Krenn

Dietel

Sers

. De novo expression of EphA2 in osteosarcoma modulates activation of the mitogenic signalling pathway. Histopathology. 2010 Dec; 57(6): 836-50.

15.

Kobayashi

Masuda

Nakayama

Ichikawa

Satow

Shitashige

Honda

Yamaguchi

Shoji

Tochigi

Morioka

Toyama

Hirohashi

Kawai

Yamada

. Reduced argininosuccinate synthetase is a predictive biomarker for the development of pulmonary metastasis in patients with osteosarcoma. Mol Cancer Ther. 2010 Mar; 9(3): 535-44.

16.

Odagiri

Kadomatsu

Endo

Masuda

Morioka

Fukuhara

Miyamoto

Kobayashi

Miyata

Aoi

Horiguchi

Nishimura

Terada

Yakushiji

Manabe

Mochizuki

Mizuta

Oike

. The secreted protein ANGPTL2 promotes metastasis of osteosarcoma cells through integrin α5β1, p38 MAPK, and matrix metalloproteinases. Sci Signal. 2014 Jan 21; 7(309): ra7.

17.

Shi

Zhou

Pan

Wei

Shi

Yao

Kang

Feng

. Exploring the key genes and pathways of osteosarcoma with pulmonary metastasis using a gene expression microarray. Mol Med Rep. 2017 Nov; 16(5): 7423-7431.

18.

Endo-Munoz

Cumming

Rickwood

Wilson

Cueva

Strutton

Cassady

Evdokiou

Sommerville

Dickinson

Guminski

Saunders

. Loss of osteoclasts contributes to development of osteosarcoma pulmonary metastases. Cancer Res. 2010 Sep 15; 70(18): 7063-72.

19.

Zhou

Zhang

Liu

Song

Shao

. LncRNA LINC00588 suppresses the progression of osteosarcoma by acting as a ceRNA for miRNA-1972. Front Pharmacol. 2020 Mar 24; 11: 255.

20.

Zhang

Guo

Zhang

Sun

Kang

Song

Liu

Lei

Wang

Guan

. Co-expression network analysis identified gene signatures in osteosarcoma as a predictive tool for lung metastasis and survival. J Cancer. 2019 Jun 9; 10(16): 3706-3716.

21.

Mohamed

FEA

Khalil

EZI

Toni

NDM

. Caveolin-1 expression together with VEGF can be a predictor for lung metastasis and poor prognosis in osteosarcoma. Pathol Oncol Res. 2020 Jul; 26(3): 1787-1795.

22.

Rutkowski

Kamińska

Kowalska

Ruka

Steffen

. Cytokine and cytokine receptor serum levels in adult bone sarcoma patients: Correlations with local tumor extent and prognosis. J Surg Oncol. 2003 Nov; 84(3): 151-9.

23.

Liu

Cheng

Santiago

Raeder

Zhang

Isabella

Yang

Semaan

Chen

Fox

Gray

Monahan

Schlegel

Beroukhim

Mills

Zhao

. Oncogenic PIK3CA-driven mammary tumors frequently recur via PI3K pathway-dependent and PI3K pathway-independent mechanisms. Nat Med. 2011 Aug 7; 17(9): 1116-20.

24.

Vogelstein

Papadopoulos

Velculescu

Zhou

Diaz

Jr Kinzler

. Cancer genome landscapes. Science. 2013 Mar 29; 339(6127): 1546-58.

25.

Kushlinskii

Fridman

Braga

. Molecular mechanisms and microRNAs in osteosarcoma pathogenesis. Biochemistry (Mosc). 2016 Apr; 81(4): 315-28.

26.

Zhou

. Focal adhesion kinase promotes progression and predicts poor clinical outcomes in patients with osteosarcoma. Oncol Lett. 2018 May; 15(5): 6225-6232.

27.

Wang

Resetkova

Wang

Correa

Hofstetter

Swisher

Ajani

Rashid

Hamilton

Albarracin

. Expression of annexin A1 in esophageal and esophagogastric junction adenocarcinomas: Association with poor outcome. Clin Cancer Res. 2006 Aug 1; 12(15): 4598-604.

28.

Wang

Chen

Wang

Pan

Xie

. Tissue microarray analysis reveals strong clinical evidence for a close association between loss of annexin A1 expression and nodal metastasis in gastric cancer. Clin Exp Metastasis. 2008; 25(7): 695-702.

29.

Bist

Leow

Phua

Shu

Zhuang

Loh

Nguyen

Zhou

Hooi

Lim

. Annexin-1 interacts with NEMO and RIP1 to constitutively activate IKK complex and NF-κB: Implication in breast cancer metastasis. Oncogene. 2011 Jul 14; 30(28): 3174-85.

30.

Liu

Huang

Zeng

Zhou

Wang

Ouyang

Cheng

. Expression of the Annexin A1 gene is associated with suppression of growth, invasion and metastasis of nasopharyngeal carcinoma. Mol Med Rep. 2014 Dec; 10(6): 3059-67.

31.

Boudhraa

Bouchon

Viallard

D’Incan

Degoul

. Annexin A1 localization and its relevance to cancer. Clin Sci (Lond). 2016 Feb; 130(4): 205-20.

32.

Foo

Yap

Cui

Lim

LHK

. Annexin-A1 – a blessing or a curse in cancer? Trends Mol Med. 2019 Apr; 25(4): 315-327.

33.

Zhang

Wang

Huang

Zheng

Cheng

. Annexin A1: A double-edged sword as novel cancer biomarker. Clin Chim Acta. 2020 May; 504: 36-42.

34.

Luker

. Functions of CXCL12 and CXCR4 in breast cancer. Cancer Lett. 2006 Jul 8; 238(1): 30-41.

35.

Xue

Mao

Ren

Chu

. Inhibition of CXCL12/CXCR4 axis as a potential targeted therapy of advanced gastric carcinoma. Cancer Med. 2017 Jun; 6(6): 1424-1436.

36.

García-Cuesta

Santiago

Vallejo-Díaz

Juarranz

Rodríguez-Frade

Mellado

. The role of the CXCL12/CXCR4/ACKR3 axis in autoimmune diseases. Front Endocrinol (Lausanne). 2019 Aug 27; 10: 585.

37.

Yazdani

Mousavi

Moradabadi

Hassanshahi

. Significance of CXCL12/CXCR4 ligand/receptor axis in various aspects of acute myeloid leukemia. Cancer Manag Res. 2020 Mar 24; 12: 2155-2165.

38.

Perissinotto

Cavalloni

Leone

Fonsato

Mitola

Grignani

Surrenti

Sangiolo

Bussolino

Piacibello

Aglietta

. Involvement of chemokine receptor 4/stromal cell-derived factor 1 system during osteosarcoma tumor progression. Clin Cancer Res. 2005 Jan 15; 11(2 Pt 1): 490-7.

39.

Oda

Yamamoto

Tamiya

Matsuda

Tanaka

Yokoyama

Iwamoto

Tsuneyoshi

. CXCR4 and VEGF expression in the primary site and the metastatic site of human osteosarcoma: Analysis within a group of patients, all of whom developed lung metastasis. Mod Pathol. 2006 May; 19(5): 738-45.

40.

Kim

Lee

Midura

Yeung

Mendoza

Hong

Ren

Wong

Korz

Merzouk

Salari

Zhang

Hwang

Khanna

Helman

. Inhibition of the CXCR4/CXCL12 chemokine pathway reduces the development of murine pulmonary metastases. Clin Exp Metastasis. 2008; 25(3): 201-11.

41.

Goguet-Surmenian

Richard-Fiardo

Guillemot

Benchetrit

Gomez-Brouchet

Buzzo

Karimdjee-Soilihi

Alemanno

Michiels

Schmid-Alliana

Schmid-Antomarchi

. CXCR7-mediated progression of osteosarcoma in the lungs. Br J Cancer. 2013 Sep 17; 109(6): 1579-85.

42.

Wang

Xue

Lin

Wang

Lin

Huang

Pan

Liu

Yan

Sun

Teng

Wang

Zhou

Chen

Poznansky

. Epigenetic regulation of CXCL12 plays a critical role in mediating tumor progression and the immune response in osteosarcoma. Cancer Res. 2018 Jul 15; 78(14): 3938-3953.

43.

Baumhoer

Smida

Zillmer

Rosemann

Atkinson

Nelson

Jundt

, von Luettichau

Nathrath

. Strong expression of CXCL12 is associated with a favorable outcome in osteosarcoma. Mod Pathol. 2012 Apr; 25(4): 522-8.

44.

Jung

Wang

Schneider

Sun

Koh-Paige

Osman

McCauley

Taichman

. Regulation of SDF-1 (CXCL12) production by osteoblasts; a possible mechanism for stem cell homing. Bone. 2006 Apr; 38(4): 497-508.

45.

Zhu

Liang

Huang

Doty

Boskey

. Conditional inactivation of the CXCR4 receptor in osteoprecursors reduces postnatal bone formation due to impaired osteoblast development. J Biol Chem. 2011 Jul 29; 286(30): 26794-805.

46.

Flores

Okcu

Murray

Chintagumpala

Hicks

Lau

Man

. Elevated expression of CXC chemokines in pediatric osteosarcoma patients. Cancer. 2011 Jan 1; 117(1): 207-17.

47.

Zhu

Huang

Xiao

Feng

Xiao

. Annexin A1-suppressed autophagy promotes nasopharyngeal carcinoma cell invasion and metastasis by PI3K/AKT signaling activation. Cell Death Dis. 2018 Nov 20; 9(12): 1154.

48.

Liao

Zhou

Zeng

Zuo

Wang

Yin

Hua

Cai

. The role of the CXCL12-CXCR4/CXCR7 axis in the progression and metastasis of bone sarcomas (review). Int J Mol Med. 2013 Dec; 32(6): 1239-46.

Identification of hub genes specific to pulmonary metastasis in osteosarcoma through integrated bioinformatics analysis

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1. Introduction

2. Materials and methods

2.1 Data sources

Table 1 Summary of microarray datasets

2.3 GO and KEGG analyses

2.4 Construction of PPI network

2.5 Identification of hub genes

2.6 Survival analysis

3.1 Identification of DEGs

3.2 Functional enrichment analysis

3.3 PPI network construction

Table 2 Top ten hub genes with higher degree of connectivity

5. Conclusion

Author contributions

Data availability

Ethical considerations

Funding

Footnotes

Conflict of interest

References

Table 1
Summary of microarray datasets

Table 2
Top ten hub genes with higher degree of connectivity