Prediction of précised biomarkers by expression-network-survival-based approach for breast cancer bone metastasis

doi:10.21203/rs.3.rs-5199308/v1

Background: About 70% of Breast Cancer (BCa) patients have Bone Metastasis (BM), the prediction of précised biomarkers for BM from BCa would guide focused treatment by early interventions to prevent or delay BM.

Method: In this study, the datasets (GSE103357, GSE55715, GSE2034, GSE14776 and GSE137842) were retrieved from the Gene Expression Omnibus (GEO). These datasets comprise of gene expression patterns of 232 samples of tumor cells from BCa and 84 samples of metastatic tumor cells from BM to understand molecular mechanism underlying in development of BCa-BM. Common differentially expressed genes (DEGs) were identified by performing meta-analysis implying ImaGEO. A protein-protein interaction (PPI) network was constructed from high throughput experiments using STRING (Search Tool for the Retrieval of Interacting Genes). Analysis of the PPI-Interactome and their sorted hub genes was performed using centrality parameters viz., degree, clustering coefficient, closeness, and betweenness centrality which are statistically and biologically significant plug-ins added in Cytoscape 3.9.1. To understand the likely course of the progression of BCa to BM survival analysis was performed using GEPIA2 and validation was done by TCGA (The Cancer Genome Atlas). To characterize the précised potential biomarkers the functional enrichment analysis, Pathway analysis and Gene Ontology (GO) studies was performed by Funrich 3.1.3.

Result: 90 common DEGs were acquired from our study among all 5 datasets, out of which 28 and 62 constitute down- and up-regulated genes respectively. Where among these 90 DEGs, 18 genes were showing unambiguous connections with each other. 12 genes were identified significant through the topological analysis performed by the above said centrality parameters of interactome and were also showing worse survival outcome in disease free survival analysis. Five genes (RACGAP1, PPP1CC, RAD23A, PSMD1 and RPL26L1) among the 12 genes were identified as hub genes and were validated by TCGA, 3 genes (RPL26L1, PSMD1, and RAD23A) were identified to show poorer disease-free survival, also RPL26L1 was identified to show worst overall survival across all samples. 3 genes (RACGAP1, PPP1CC, and RAD23A) were literature reviewed potential diagnostic biomarkers in cancer progression, including breast cancer bone metastasis. RPL26L1 is significant potential biomarker identified with upregulation in BCa-BM samples with worse overall survival. These reported and significantly identified genes intersect notably cell growth, skeletal muscle development, and cell communication

Conclusion: Our expression-based network-based approach successfully prioritized précised biomarkers for breast cancer bone metastasis.

Bone metastasis

Differentially expressed genes

and Network analysis

Breast cancer is one of the most common malignant tumors in gynecology showing a high global incidence with over 2,261,419 of new cases and the mortality with over 684,996 of deaths in 2020 (1). The probability of recurrence lasts for up to 20 years, even though 60–80% of recurrences happen in the first three years (2, 3). Despite the fact that surgery may cure the majority of breast cancers, 25% have a latent and sneaky nature that spreads quickly despite developing slowly. If the microenvironment is favorable, tumor cells will develop metastasis and grow in the distant location (4, 5). Recently, multiple patterns of gene expression connected to breast cancer metastatic characteristics have been discovered by genomic profiling using microarrays (6, 7). Bone is the third most frequent site of metastasis after lung and liver (Maisano et al., 2001a; Theriault & Theriault, 2012)(10). Up to 70% of skeletal bone metastasis (BM) is reported to be caused by breast cancer (11). Pain, confined movement, nerve root lesions, spinal cord compression, cranial nerve palsies, pathological fractures, hypercalcemia, and restriction of bone marrow function are among the complications of bone metastasis (12, 13).

We now understand that osteotropism, the term for cancer that metastasizes to bone, involves a series of sequential events, involving the acquisition of unique genetic traits by tumor cells in order for them to: (1) separate from the main tumor; (2) infiltrate the bone; and (3) establish a home inside the bone niche. Nevertheless, despite significant progress in cancer treatment, the molecular mechanisms underlying metastasis remain unexplained. Furthermore, only a limited subset of patients benefits from adjuvant treatment with denosumab or bisphosphonates (14–16). As a result, several teams have started employing molecular biology techniques to try and decipher BM processes (17). Many interactions between the cancer cells and the targeted organ are necessary for tumor cells to proliferate in remote locations. Recent research employing animal models and human breast cancer cell lines has demonstrated that bone metastasis may not happen unless certain genes are expressed (18). Many of these genes are still unknown therefore, it is possible that the identification of molecular gene signatures may help identify targeted medicines in addition to potentially predicting tumors with a high propensity for metastasis in the near future (19).

The present study aimed to explore target genes or biomarkers involved for breast cancer bone metastasis (BCa-BM) by applying network studies to meta-analyzed expression data. We performed meta-analysis of five breast cancer gene expression data in order to identify differentially expressed genes (DEGs) between primary and metastasized tumors. Subsequently, we carried out functional and pathway enrichment analyses on DEGs, which were then combined to build a network of protein-protein interactions (PPIs) and filter hub genes. Key biomarkers for BCa-BM were identified using additional functional enrichment analysis, pathway analysis, and Gene Ontology (GO) examinations of the significant genes. Further survival analysis was performed and significant genes were validated by external dataset from TCGA (The Cancer Genome Atlas).

Data retrieval:

The following query was searched: using the search term "(bone metastasis AND breast cancer) AND "Homo sapiens"[porgn:_txid9606]", in the Gene expression omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) (20) to discover the gene expression datasets. (4). Expression datasets with primary tumor (breast cancer without metastasis) and tumors with BM (breast cancer metastasized to bone) were filtered amongst the available 300 entries. The datasets GSE103357, GSE55715, GSE2034, GSE14776, and GSE137842 were acquired using the different platforms of GEO (Table 1). Expression value of all datasets was analysed by limma package of R (Supplementary file 1). The ImaGEO tool (21) was utilised to execute the preliminary differential expression analysis, which involved batch effect correction, normalisation, and background correction.

Table 1: Parameters of different gene IDs for primary breast cancer and bone metastases that were obtained from the GEO database

Gene ID’s	Platform	Sample size	Primary tumor	Metastasis
GSE103357	GPL6974	5	2	3
GSE55715	GPL6947	8	2	3
GSE2034	GPL96	286	217	69
GSE14776	GPL6883	14	8	6
GSE137842	GPL570	10	3	3

Integrated analysis:

Meta analysis is a statistical method for integrating the findings of several studies on the same subject, and it's becoming more and more common for resolving conflicts in clinical research (22). For the meta-analysis of the gene expression data, we utilised ImaGEO software with the effect size approach and the default settings. When necessary, the tool converts expression values to the logarithmic scale, adds distinct Entrez Gene identifiers to the probe identifiers, combines the data, and performs data quality control assessments. Using the average expression values in the corresponding primary or metastasis group, the tool imputes missing values for the remaining genes and further computes median values for duplicate gene expression profiles in each dataset. It also filters out genes with missing values in more than 10% of samples. Using MetaDE.ES from the Meta DE package, we determined the DEGs. This approach utilized three statistical factors to examine the heterogeneity of gene expression value: Q-value, τ² and Qpval . Next, we used the p-value to test for gene expression differences between the main and metastasized groups. The cutoffs were set at τ² = 0, Qpval >0.05, and p < 0.05 to guarantee the homogeneity of featured genes. False discovery rate (FDR) p-value < 0.05 and log2fold change >2 were the requirements for DEGs. As a result, unlike the widely used limma, which selects DEGs based on p-value and fold-change thresholds, the MetaDE package first performs heterogeneity tests to ascertain whether genuine differences, as opposed to variation based on chance alone, underlie the results of the studies (heterogeneity) before selecting DEGs successively (23, 24).

Functional enrichment analysis:

STRING (Search Tool for the Retrieval of Interacting Genes) database which is accessible online (http://www.bork.embl-heidelberg.de/STRING/), (25) has been used in our study for construction of protein-protein interaction network (PPIN). Among the initial approaches, STRING aims to stand out by prioritizing good coverage, usability, and a reliable scoring method. It presently has 2,61,033 orthologs, along with 89 completely sequenced genomes which are updated continually. The database, predicts functional connections with an estimated degree of accuracy. The interactions in string analysis include direct (physical) and indirect (functional) associations. Coverage (relevant to several thousand of genome-sequenced organisms), richness of evidence sources (such automated content mining), and usability factors (such as customization, enrichment detection and programmatic access) are the focused areas for STRING (26). In this study, identified DEGs were further used to construct a PPIN using STRING version 12.0. The cut-off criterion for identifying a meaningful interaction in this PPIN was a medium confidence interaction score of 0.400.

Network and molecular profile analysis:

Open-source software called Cytoscape3.9.1 (27) was used to integrate, visualize, and analyze biological networks. The incorporation of Cytoscape into network analysis tool promises greater usability and wide spread adoption of our method (28). A Cytoscape plug in, cytoHubba was used for the further study in analyzing network and molecular profile. Eleven topological analysis techniques are available through cytoHubba, including six centralities based on shortest paths (bottleneck, EcCentricity, closeness, radiality, betweenness, and stress), Maximum Clique Centrality (MCC), Maximum Neighborhood Component, Degree, Edge Percolated Component, and Density of Maximum Neighborhood Component (29). A Cytoscape plug-in called Network Analyzer (30) was used for the genes showing direct interactions in STRING. It adds node properties for the outcomes and provides helpful visualization settings to display and export the generated distributions. A few centrality measures, including degree, clustering coefficient, closeness, and betweenness centrality were indicated. On the basis of shortest path using betweenness centrality and MCC parameters top hub genes were fetched via cytoscape plug-in cytoHubba. MCC surpasses the other approaches among the 11 techniques. In both high-degree and low-degree proteins, MCC captures more crucial proteins in the top ranked list (31). To increase the specificity and sensitivity, MCC is the best method to discover featured nodes. Consider a node v, the MCC of v is defined as MCC(v)=∑C∈S(v)(|C|−1)! where S(v) represents the set of maximum cliques that contain v. The sum of all positive integers below |C| is (|C|-1)! MCC(v) is equal to the node's degree if there is no edge separating its neighbors at node v.

Survival analysis and TCGA validation:

The Kaplan–Meier method was used to plot survival curves. The disease-free survival and overall survival (OS) of patients with BCa in BRCA (breast invasive carcinoma) were compared using the hazard ratio (HR) with 95% confidence intervals (CIs) and log rank P-value. Hub genes were evaluated for their prognostic importance using the Gene Expression Profiling Interactive Analysis (GEPIA2; http://gepia.cancer-pku.cn). To determine which group has the worst survival among the given Kaplan-Meier survival curves, we can look at the overall shape and the position of the survival curves. The group with the lowest survival rates at the end of the study period would be considered to have the worst survival (32). Further validation of identified hub genes is performed by taking external dataset of BRCA from TCGA, which contains 291 normal samples and 1085 tumor samples. Expression values of hub genes were plotted using Gene Expression Profiling Interactive Analysis.

Functional characterization and pathway analysis:

Functional characterization, pathway analysis, and gene ontology (GO) studies were performed by using FunRich (http://www.funrich.org), KEGG and GeneCard. A functional enrichment study may be carried out using FunRich applying background databases that have been combined from various genomic and proteomic resources (>1.5 million annotations) (33). The background tool is very important in any analysis tool so, it is updated continuously to give the appropriate results (34). It is a standalone program that combines simplicity of use with customized databases, open access, and graphical representations (Venn, pie charts, bar graphs, column, heatmap and doughnuts) (35). These programs use a similar approach to comprehensively map the list of genes and proteins to a unique database and offer a list of overrepresented and underrepresented classes of annotated items. When integrating and interpreting enormous datasets produced by high-throughput experimental techniques, as well as for determining their useful values, KEGG may be utilized as an important information source (36). Gene card combines the gene data from other large public resources, such as NCBI, UniPortKB, HGNC, ENSEMBL as well as many other public sources, on all known and anticipated human genes, it has supplied brief function, disease, genome, proteome and transcriptome data (37)

Study design:

Six steps make up the study's design, which is shown in a flowchart in Fig. 1. First, we combined and analyzed the five GEO datasets by Meta-analysis using ImaGEO to produce DEGs (Step 1). Then, the DEGs were used to construct a gene-gene functional interaction network in String database (Step 2). Next, the network was analyzed by different Cytoscape plug-ins (Step 3). Genes were sorted considering four centrality parameters and hub genes were identified. (Step 4). Next, survival analysis was performed on identified hub genes and these genes were validated by TCGA database (Step 5). At last Functional characterization, Pathway analysis and GO studies were done with the help of FunRich, KEGG and GeneCards (Step 6).

Differentially Expressed Genes (DEGs) between primary and bone metastasized tumors:

The ImaGEO algorithm merged and normalized the data after further annotating the probes with gene IDs to obtain the DEGs. The results of the quality control tests performed by the ImaGEO tool for the five gene expression datasets are given in the (Table 2), which shows good quality of data. We acquired 90 DEGs from the tool ImaGEO (Supplementary file 2). The expression of the genes is shown using a heatmap, which is an image representation of all the genes (Fig. 2) and the box plots of the five datasets showing expression values is represented (Fig. 3).

Table 2

Dataset quality control results generated by the ImaGEO tool
ID	Samples preQC	Samples postQC	PASS_QC
GSE103357	286	286	YES
GSE55715	14	14	YES
GSE2034	5	5	YES
GSE14776	5	5	YES
GSE137842	6	6	YES

Functional enrichment analysis:

The prior step finalizes 90 DEGs which undergoes functional enrichment analysis. To understand the relationships between DEGs at the protein level, a protein-protein interaction network (PPIN) (38) was constructed. In the initial stage of STRING database, it has been observed that 18 genes were showing direct connection with each other (Fig. 4).

Network and molecular profile analysis:

Total 12 genes (RAD23A, PPP1CC, RPL26L1, COL5A1, CHEK2, CCT6A, PSMD1, CRK, MLH3, RACGAP1, NEK2 and COX7C) were identified to cross mentioned centrality parameters (Supplementary file 3). A venn diagram constructed by https://bioinformatics.psb.ugent.be/webtools/Venn/ to show number of genes identified to crossing all centrality parameters (Fig. 5). Out of these 5 genes (RACGAP1, PPP1CC, RAD23A, PSMD1 and RPL26L1) were identified as hub genes on the basis of ranking in MCC plug-In (Fig. 6).

Survival Analysis:

Based on the Kaplan-Meier curves provided for Disease-Free Survival (DFS), we can say that the DFS is not good in several of the genes. Particularly DFS is not good in the high expression groups for the genes RPL26L1, PSMD1, and RAD23A (Fig. 7). Therefore RPL26L1, PSMD1, and RAD23A shows prognostics significance in DFS. In addition, RPL26L1 (Plot A) shows prognostic significance in overall survival (Fig. 8), as the high group has a noticeably lower survival rate compared to the other plots. All five genes (RACGAP1, PPP1CC, RAD23A, PSMD1 and RPL26L1) has shown significant difference in expression level between tumor and normal samples for BRCA dataset in TCGA (Fig. 9).

Pathway Analysis and Functional Enrichment Analysis:

Molecular function study revealed that the majority of these genes-controlled protein binding and activities, as shown in Table 3. All of these genes appeared to have a specific role in cell growth, skeletal muscle development, and cell communication.

Table 3

Gene Ontology: Biological Pathways and Functions of Identified Hub Genes.
Gene Name	Molecular Function	Biological Process	Cellular Component	Biological Pathway
RACGAP1	GTPase activator activity	Cell communication, Signal transduction	Microtubule, Nucleus and Exosomes	Aurora B signaling, Signaling by Aurora kinases and Kinesins
PPP1CC	Protein serine/threonine phosphatase activity	Signal transduction	Microtubule, Protein complex, Nucleus and Cytoplasm	Aurora B signaling, Mitotic M-M/G1 phases, DNA Replication and Insulin-mediated glucose transport
RAD23A	DNA binding	DNA repair	Nucleus	Nucleotide excision repair and Protein processing in endoplasmic reticulum
PSMD1	Ubiquitin-specific protease activity	Protein metabolism	Proteasome regulatory particle, Cytosol, Nucleus, Lysosome and Cytoplasm	Metabolism of mRNA, Cell Cycle, Mitotic and Metabolism of RNA
RPL26L1	RNA binding and Structural constituent of ribosome	Biological_process unknown	Ribosome, Cytosol, Nucleolus and Exosomes	Viral mRNA Translation, Eukaryotic Translation Termination and Metabolism of RNA

This study aimed to identify precise biomarkers for breast cancer bone metastasis (BCa-BM) using a comprehensive bioinformatics approach. Utilizing five datasets from the Gene Expression Omnibus (GEO) (39), comprising 232 BCa tumor samples and 84 BM metastatic tumor samples, the analysis identified 90 common differentially expressed genes (DEGs). Among these, 28 were down-regulated and 62 were up-regulated in BM samples compared to BCa samples. Notably, 18 genes exhibited significant interactions within a protein-protein interaction (PPI) network constructed using STRING (40), highlighting their central roles in BCa-BM progression. The identification of hub genes such as RACGAP1, PPP1CC, RAD23A, PSMD1, and RPL26L1 underscores their potential significance in BCa metastasis to bone. These genes, validated through topological analysis in Cytoscape 3.9.1 (41), are implicated in critical biological processes like cell growth, skeletal muscle development, and cell communication. Importantly, RPL26L1 emerged as a standout biomarker with upregulation in BCa-BM samples correlating significantly with worse overall survival, as validated by TCGA data.

Survival analysis using GEPIA2 and TCGA(42) data further supported the prognostic relevance of these biomarkers, particularly highlighting RPL26L1, PSMD1, and RAD23A for their association with poorer disease-free survival outcomes. Literature review also substantiated RACGAP1, PPP1CC, and RAD23A as potential diagnostic biomarkers in cancer progression, including breast cancer bone metastasis, reinforcing their roles in metastatic processes (43–46). GTPase-activating protein (GAP) is encoded by the protein-coding gene RACGAP1 (Rac GTPase-activating protein 1). According to Zhou et al. (2021), there was a high expression of RACGAP in breast cancer, which was linked to poor overall survival and several subtypes of the disease (43). Numerous cellular processes, including as cytokinesis, transformation, invasive migration, and metastasis, have been linked to the vital protein RACGAP1. Numerous investigations have examined RACGAP1's function in various human cancer types (44). Belonging to the Serine/threonine-protein phosphatase family, protein phosphatase-1 (PP1) is involved in several physiological processes, such as glycogen metabolism, cellular proliferation, and cell receptor modulation. Human mammary tissues express the protein phosphatase-1 catalytic units, PPP1CA, PPP1CB, and PPP1CC; elevated levels of these units represent a positive clinical signal for patient survival and a lower risk of bone metastases (45). Rad23 is involved in several key biological processes, including apoptosis, development, DNA repair, cell cycle regulation, and carcinogenesis (47). Additionally, it was discovered that the expression profiles of the metastasis-specific protein RAD23B correlated with the overall survival of invasive BC patients and were consistent with their gene expression levels in BRCA patients. These findings raise the possibility that these proteins could serve as potential exosome markers for BC metastasis (46).

Functional enrichment and pathway analyses revealed that these DEGs are predominantly involved in pathways related to cell cycle regulation, skeletal muscle development, and cell communication. Disruptions in these pathways may facilitate the metastatic cascade, providing potential targets for therapeutic interventions aimed at delaying or preventing bone metastasis in breast cancer (48).

The identification of precise biomarkers such as RACGAP1, PPP1CC, RAD23A, PSMD1, and particularly RPL26L1, offers significant clinical implications. These biomarkers can potentially enhance the accuracy of prognostic models, guiding personalized treatment strategies for breast cancer patients at risk of bone metastasis. Early intervention targeting these biomarkers could improve patient outcomes by mitigating the progression of metastasis.

The integration of bioinformatics tools and comprehensive datasets from GEO and TCGA underscores the robustness of the findings (49). However, the reliance on gene expression data necessitates further experimental validation to elucidate the functional roles of these biomarkers in vivo and to explore their potential as therapeutic targets (50).

Despite its strengths, this study is limited by its reliance on retrospective data analysis and bioinformatics predictions. The findings require validation through experimental studies to confirm the functional roles of identified biomarkers in BCa-BM progression. Additionally, the study's focus on a subset of DEGs may overlook other relevant genes and pathways contributing to metastasis.

Future research should consider expanding the analysis to include broader genomic and clinical data sets, as well as experimental validation in preclinical models and clinical samples. Such efforts will help translate these bioinformatics insights into clinically actionable strategies for managing breast cancer metastasis.

Future research should prioritize experimental validation of RACGAP1, PPP1CC, RAD23A, PSMD1, and RPL26L1 using cell culture models and animal studies. Investigating the molecular mechanisms underlying their roles in BCa-BM progression, particularly focusing on interactions within the identified pathways, will provide deeper insights into metastatic processes. Moreover, exploring targeted therapies aimed at modulating the cell cycle, skeletal muscle development, and cell communication pathways could yield novel strategies for preventing or treating bone metastasis in breast cancer patients. Integrating multi-omics approaches and functional assays will further enhance our understanding of these biomarkers' clinical relevance.

In conclusion, this study's expression-based network approach successfully prioritized precise biomarkers for breast cancer bone metastasis. The identification of hub genes and their association with clinical outcomes highlights their potential as prognostic markers and therapeutic targets. These findings pave the way for advancing personalized medicine in breast cancer management, offering hope for improved outcomes through targeted interventions aimed at mitigating bone metastasis. Further research and validation are essential to translate these discoveries into clinical practice effectively.

Ethical Approval: NA

Consent to Participate: All authors confirm that they have fully participated in this research and approve the manuscript's contents. Each author has contributed to the study in line with the authorship guidelines and has reviewed and agreed to submit this manuscript.

Consent to Publish: All authors provide consent to publish this manuscript in Applied Biochemistry and Biotechnology. The content is original, has not been previously published, and all authors approve its submission and eventual publication. No ethical concerns exist with the data or manuscript contents.

Author contribution:

Mahima Bhardwaj: Conceived the study and designed the experiments. Primarily responsible for overseeing the project and drafting the initial manuscript, also performed critical revisions and approved the final version of the manuscript.

K. Abraham Peele: Involved in data analysis and interpretation. Additionally, contributed significantly to the writing of the methods and results sections of the manuscript.

Sachidanand Singh: Assisted in data interpretation and contributed to the writing and editing of the discussion section, also reviewed the final manuscript for important intellectual content and approved the final version for publication.

All authors have read and approved the final manuscript.

Data availability: Data is online available in NCBI database, also supplementary files are provided with manuscript.

Funding: NA

Competing Interests: There are no financial or non-financial conflicts of interest among the authors associated with this submission. We confirm that the content of this manuscript is original, and we have appropriately cited the work of others where necessary. We comply with the ethical guidelines provided by the Journal and have adhered to the principles of transparency and accuracy throughout the research process.

Should our manuscript be accepted for publication, we understand that this declaration of interest will be made public to ensure transparency and maintain the credibility of the scientific community.

Sung H, Ferlay J, Siegel RL, et al (2021) Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 71:209–249
Chambers AF, Groom AC, and MacDonald IC (2002) Dissemination and growth of cancer cells in metastatic sites. Nat Rev Cancer 2:563–572
Bussard KM, Gay C V., and Mastro AM (2008) The bone microenvironment in metastasis; what is special about bone? 27:41–55
Vijver MJ van de, He YD, ’t Veer LJ van, et al (2002) A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999–2009
Coleman RE (2001) Metastatic bone disease: clinical features, pathophysiology and treatment strategies. Cancer Treat Rev 27:165–176
Ramaswamy S, Ross KN, Lander ES, et al (2003) A molecular signature of metastasis in primary solid tumors. Nat Genet 33:49–54
Kang Y, Siegel PM, Shu W, et al (2003) A multigenic program mediating breast cancer metastasis to bone. Cancer Cell 3:537–549
Maisano R, Pergolizzi S, and Cascinu S (2001) Novel therapeutic approaches to cancer patients with bone metastasis. Crit Rev Oncol Hematol 40:239–250
Theriault RL and Theriault RL (2012) Biology of bone metastases. Cancer Control 19:92–101
Minn AJ, Gupta GP, Siegel PM, et al (2005) Genes that mediate breast cancer metastasis to lung. Nature 436:518–524
Maisano R, Pergolizzi S, and Cascinu S (2001) Novel therapeutic approaches to cancer patients with bone metastasis. Crit Rev Oncol Hematol 40:239–250
Theriault RL and Theriault RL (2012) Biology of bone metastases. Cancer Control 19:92–101
Bone Metastasis - PubMed, https://pubmed.ncbi.nlm.nih.gov/29939688/
Jacobs C, Amir E, Paterson A, et al (2015) Are adjuvant bisphosphonates now standard of care of women with early stage breast cancer? A debate from the Canadian Bone and the Oncologist New Updates meeting. J Bone Oncol 4:54–58
Gnant M, Mlineritsch B, Stoeger H, et al (2015) Zoledronic acid combined with adjuvant endocrine therapy of tamoxifen versus anastrozol plus ovarian function suppression in premenopausal early breast cancer: final analysis of the Austrian Breast and Colorectal Cancer Study Group Trial 12. 26:313–320
Paterson AH, Anderson SJ, Lembersky BC, et al (2012) Oral clodronate for adjuvant treatment of operable breast cancer (National Surgical Adjuvant Breast and Bowel Project protocol B-34): a multicentre, placebo-controlled, randomised trial. Lancet Oncol 13:734–742
Kingsley LA, Fournier PGJ, Chirgwin JM, et al (2007) Molecular Biology of Bone Metastasis. Mol Cancer Ther 6:2609–2617
Kominsky SL and Davidson NE (2006) A “bone” fide predictor of metastasis? Predicting breast cancer metastasis to bone. J Clin Oncol 24:2227–2229
Bhardwaj M, Sri T, Krupanidhi S, et al (2024) Identification of potential biomarkers for bone metastasis using human cancer metastasis database. Int J Health Sci (Qassim) 18:6–14
Barrett T, Suzek TO, Troup DB, et al (2005) NCBI GEO: mining millions of expression profiles--database and tools. Nucleic Acids Res 33
Toro-Domínguez D, Martorell-Marugán J, López-Domínguez R, et al (2019) ImaGEO: integrative gene expression meta-analysis from GEO database. 35:880–882
Walker E, Hernandez A V., and Kattan MW (2008) Meta-analysis: Its strengths and limitations. Cleve Clin J Med 75:431–439
Wei L, Wang Q, Zhang Y, et al (2018) Integrated analysis of microarray data to identify the genes critical for the rupture of intracranial aneurysm. Oncol Lett
Albaradei S, Uludag M, Thafar MA, et al (2021) Predicting Bone Metastasis Using Gene Expression-Based Machine Learning Models. Front Genet 12
Mering C von, Huynen M, Jaeggi D, et al (2003) STRING: a database of predicted functional associations between proteins. Nucleic Acids Res 31:258–261
Szklarczyk D, Gable AL, Nastou KC, et al (2021) The STRING database in 2021: customizable protein–protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res 49:D605–D612
Shannon P, Markiel A, Ozier O, et al (2003) Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Res 13:2498–2504
Saito R, Smoot ME, Ono K, et al (2012) A travel guide to Cytoscape plugins. Nat Methods 9:1069–1076
Chin C-H, Chen S-H, Wu H-H, et al (2014) cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC Syst Biol 8:S11
Assenov Y, Ramírez F, Schelhorn S-E, et al (2008) Computing topological parameters of biological networks. 24:282–284
Chin C-H, Chen S-H, Wu H-H, et al (2014) cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC Syst Biol 8:S11
Zheng S, Yang L, Dai Y, et al (2019) Screening and Survival Analysis of Hub Genes in Gastric Cancer Based on Bioinformatics. 26:1316–1325
Pathan M, Keerthikumar S, Ang C, et al (2015) FunRich: An open access standalone functional enrichment and interaction network analysis tool. Proteomics 15:2597–2601
Wadi L, Meyer M, Weiser J, et al (2016) Impact of outdated gene annotations on pathway enrichment analysis. Nat Methods 13:705–706
Benito‐Martin A and Peinado H (2015) FunRich proteomics software analysis, let the fun begin! Proteomics 15:2555–2556
Kanehisa M, Goto S, Sato Y, et al (2012) KEGG for integration and interpretation of large-scale molecular data sets. Nucleic Acids Res 40:D109–D114
Safran M, Dalah I, Alexander J, et al (2010) GeneCards Version 3: the human gene integrator. Database 2010:baq020–baq020
George G, Singh S, Lokappa SB, et al (2019) Gene co-expression network analysis for identifying genetic markers in Parkinson’s disease - a three-way comparative approach. Genomics 111:819–830
Clough E, Barrett T, Wilhite SE, et al (2024) NCBI GEO: archive for gene expression and epigenomics data sets: 23-year update. Nucleic Acids Res 52:D138–D144
Franz M, Lopes CT, Fong D, et al (2023) Cytoscape.js 2023 update: a graph theory library for visualization and analysis. 39
Franz M, Lopes CT, Fong D, et al (2023) Cytoscape.js 2023 update: a graph theory library for visualization and analysis. 39
Filippi A and Mocanu M-M (2023) Mining TCGA Database for Genes with Prognostic Value in Breast Cancer. Int J Mol Sci 24:1622
Zhou Q, Liu X, Lv M, et al (2021) Genes That Predict Poor Prognosis in Breast Cancer via Bioinformatical Analysis. Biomed Res Int 2021:1–8
Eid RA, Soltan MA, Eldeen MA, et al (2022) Assessment of RACGAP1 as a Prognostic and Immunological Biomarker in Multiple Human Tumors: A Multiomics Analysis. Int J Mol Sci 23:14102
Dong X, Martin TA, Dou Q, et al (2023) Abstract P2-26-11: Protein phosphatase 1 catalytic units, PPP1CA, PPP1CB and PPP1CC in breast cancer. Cancer Res 83:P2-26-11-P2-26–11
Shen S, Tu C, Shen H, et al (2023) Comparative Proteomics Analysis of Exosomes Identifies Key Pathways and Protein Markers Related to Breast Cancer Metastasis. Int J Mol Sci 24:4033
Yokoi M and Hanaoka F (2017) Two mammalian homologs of yeast Rad23, HR23A and HR23B, as multifunctional proteins. Gene 597:1–9
Song X, Wei C, and Li X (2022) The Signaling Pathways Associated With Breast Cancer Bone Metastasis. Front Oncol 12
Fan W, Chen X, Li R, et al (2023) A prognostic risk model for ovarian cancer based on gene expression profiles from gene expression omnibus database. Biochem Genet 61:138–150
Zhao W, Chang Y, Wu Z, et al (2023) Identification of PIMREG as a novel prognostic signature in breast cancer via integrated bioinformatics analysis and experimental validation. PeerJ 11:e15703

Prediction of précised biomarkers by expression-network-survival-based approach for breast cancer bone metastasis

Status:

Version 1

Abstract

Figures

Introduction

Methodology

Results

Study design:

Differentially Expressed Genes (DEGs) between primary and bone metastasized tumors:

Functional enrichment analysis:

Network and molecular profile analysis:

Survival Analysis:

Pathway Analysis and Functional Enrichment Analysis:

Discussion

Statements & Declarations

References

Supplementary Files

Status:

Version 1