Abstract

Background

   Long non-coding RNAs (lncRNAs) play crucial roles in the progression of
   breast cancer (BC). The lncRNA small nuclear RNA host gene 16 (SNHG16),
   represents a lncRNA associated with tumor development. This study
   focuses on SNHG16 and its regulatory role in BC progression, validated
   through microarray analysis and in vitro experiments.

Method

   The GEO datasets [41]GSE65194 and [42]GSE61304 were used to identify
   differentially expressed lncRNAs, [43]GSE41922 and [44]GSE45666 for
   differentially expressed miRNAs, and [45]GSE29431 and [46]GSE42568 for
   differentially expressed mRNAs. The TCGA database was used to validate
   the genes identified in the previous analyses. The regulatory pathways
   of SNHG16 were identified through differential gene analysis, target
   gene prediction, functional enrichment analysis, and protein-protein
   interaction network analysis. RT-qPCR, Western blot, CCK-8, cell
   migration, and cell knockdown experiments were performed for
   validation.

Results

   Our study found that elevated expression levels of SNHG16 were
   associated with BC cell proliferation and poor prognosis. According to
   an in vitro knockdown assay, SNHG16 silencing inhibits the progression
   of BC. SNHG16 was found to positively regulate AURKA levels as a
   competitive sponge for hsa-let-7b-5p.

Conclusion

   The activity of the lncRNA SNHG16 can be triggered via the
   hsa-let-7b-5p/AURKA axes. These findings shed light on the novel
   molecular mechanisms underlying BC progression.

   Keywords: Breast cancer, SNHG16, ceRNA network, Proliferation,
   Migration

Highlights

     * •
       SNHG16 is significantly differentially expressed in breast cancer
       through TCGA and GEO databases.
     * •
       The lncRNA SNHG16 is involved in the ceRNA regulatory network.
     * •
       qRT-PCR showed that SNHG16 was highly expressed in breast cancer
       cell lines.
     * •
       Knockdown of SNHG16 decreased the proliferation and metastasis of
       breast cancer cells.

1. Introduction

   Breast cancer (BC) is a malignant tumor originating in the epithelial
   tissue of the breast, commonly causing female mortality and increasing
   incidence among young women [[47]1,[48]2]. Advances in treatment have
   significantly reduced mortality rates, but the primary cause of death
   remains metastasis rather than the primary tumor [[49]3]. Research
   shows that over 90 % of breast cancer-related deaths are due to
   metastasis, highlighting its critical role in mortality [[50]4,[51]5].
   Therefore, identifying molecular biomarkers for early detection and
   understanding metastasis mechanisms are crucial for effective
   treatment.

   The complex interplay of multiple genes, factors, and molecules drives
   the development of breast cancer. Long non-coding RNAs (lncRNAs) are
   found in the nucleus or cytoplasm, they regulate gene expression via
   competing endogenous RNA (ceRNA) mechanisms. Dysregulated expression of
   lncRNAs contributes to disease progression through ceRNA regulatory
   networks [[52][6], [53][7], [54][8]]. The ceRNA hypothesis proposes
   that lncRNAs and mRNAs can competitively bind microRNAs (miRNAs)
   through shared microRNA response elements (MREs) [[55][9], [56][10],
   [57][11], [58][12]]. This interaction reduces miRNA-mediated
   suppression of mRNA, influencing biological processes such as
   proliferation, migration, and apoptosis in tumor cells [[59]13,[60]14].

   Recent studies have shown that lncRNA SNHG16 negatively regulates
   miR-218-5p expression in pancreatic cancer, thereby inhibiting cancer
   cell invasion and metastasis [[61]15,[62]16]. In breast cancer, SNHG16
   similarly acts as a molecular sponge, regulating downstream target
   genes and thus affecting migration [[63]17]. Downregulation of
   has-let-7b in ovarian cancer promotes oncogenic gene expression,
   enhancing cell proliferation and migration [[64]18]. Elevated
   miR-196a-5p levels in triple-negative breast cancer cells promote
   proliferation and migration, which decrease upon inhibition of upstream
   genes [[65]19]. Research has validated that AURKA impacts cell
   proliferation in hepatocellular, gastric, and breast cancer cells
   [[66][20], [67][21], [68][22], [69][23]]. However, whether SNHG16 acts
   as a ceRNA to regulate downstream AURKA expression in breast cancer
   remains unexplored.

   This study uses the TCGA and GEO databases to identify differentially
   expressed RNAs related to breast cancer prognosis, constructing a ceRNA
   expression network. Experimental validation of this ceRNA network's
   impact on breast cancer proliferation and invasion provides new
   biomarkers and theoretical foundations for understanding disease
   progression.

2. Materials and methods

2.1. Microarray analysis

   The Cancer Genome Atlas (TCGA) is the largest sequencing database
   dedicated to breast cancer research [[70]24]. We screened 943 eligible
   breast cancer tissue samples and 95 adjacent normal tissue samples,
   merged the data from both using Perl scripts, and compiled mRNA
   expression profiles for 60,488 Ensembl genes and miRNA expression
   profiles of 1882 genes. Genes were annotated using the Homo sapiens
   genome information file (Homo_sapiens.GRCh38.105.chr.gtf),
   distinguishing between lncRNAs and mRNAs. Meanwhile, the Gene
   Expression Omnibus (GEO) is the most comprehensive repository of
   microarray data worldwide [[71]25]. For our study, we selected the
   breast cancer gene expression microarrays [72]GSE65194 and [73]GSE61304
   to identify differentially expressed lncRNAs; [74]GSE41922 and
   [75]GSE45666 for miRNAs; and [76]GSE29431 and [77]GSE42568 for mRNAs.
   Detailed dataset information is provided in [78]Table S1.

2.2. Identification of differentially expressed genes (DEGs)

   We conducted RNA sequencing analysis using the limma package in R to
   group, deduplicate, and analyze RNA samples. DEGs were defined based on
   log[2] (fold change) > 2 or log[2] (fold change)<-2 (log[2]FC>|2|) and
   with P < 0.05.

   Visualization of DEGs was achieved using ggplot2 for volcano plots and
   pheatmap for heatmaps. Upregulated genes were marked in red, while
   downregulated genes were marked in green. We performed an intersection
   analysis of DEG results across various databases and gene chips.

2.3. Functional enrichment analysis of DEGs

   Functional characterization and enrichment analysis of DEGs were
   conducted using Gene Ontology (GO) terms, categorizing biological
   processes (BP), molecular functions (MF), and cellular components (CC).
   Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment
   analysis identified significant pathways enriched with DEGs compared to
   the whole genome background. Visualization of enrichment results was
   performed using ggplot2 in R (4.2.0).

2.4. Protein-protein interaction (PPI) network analysis of DEGs

   PPI networks illustrate interactions between proteins, providing
   insights into downstream protein expression regulation. DEGs were
   analyzed using the STRING database, and the resulting PPI network was
   visualized using Cytoscape. The CytoHubba plugin identified key
   proteins based on their degree of interaction. Additionally, core
   subnetworks were extracted using the MCODE plugin with specific
   parameters.

2.5. Prediction of target genes of DEGs

   We utilized TargetScan (TargetScan:[79]http://targetscan.org/), miRanda
   (miRanda:[80]http://microrna.org/microrna/), and Starbase (Starbase,
   [81]https://starbase.sysu.edu.cn/) to predict target genes, focusing on
   the interactions of miRNA, mRNA, and lncRNA. Predictions were refined
   by intersecting results from these platforms, ensuring accuracy. A Venn
   diagram was generated to visualize common target genes, which were
   subsequently included in further analyses. Target gene prediction
   websites.

2.6. Construction of ceRNA regulatory network

   We utilized Targetscan and miRanda to predict target binding
   relationships between miRNA, mRNA, and lncRNA. Specifically, we
   selected DEGs with Targetscan scores ≥50, and miRanda energy 0.1.
   According to the competitive endogenous RNA (ceRNA) hypothesis,
   DElncRNAs and DEmRNAs within DEGs are expected to exhibit similar
   expression trends, while DEmiRNAs should show opposite trends
   [[82]26,[83]27]. Using these principles, we constructed differential
   expression lncRNA-miRNA-mRNA ceRNA regulatory networks and visualized
   them using Cytoscape software.

2.7. Survival and expression analysis of DEGs

   We performed survival analysis using the Kaplan-Meier Plotter database
   ([84]http://kmplot.com), which evaluates gene expression across 21
   cancer types [[85]28]. This database assessed the prognosis of
   differentially expressed mRNAs, miRNAs, and lncRNAs identified in our
   study. Additionally, we utilized GEPIA (Gene Expression Profiling
   Interactive Analysis), integrating data from TCGA to analyze DEG
   expression across tissues [[86]29]. UALCAN is a TCGA-based resource for
   further expression analysis and biomarker identification [[87]30].
   Proteinatlas provided insights into gene and protein expression in
   various tissues. These databases collectively informed us about DEG
   expression disparities between breast cancer and normal tissues.

2.8. Cell culture and transfection

   The MCF-7 (SCSP-531) and MDA-MB-231 (SCSP-5043) cell lines,
   representing two prevalent and highly proliferative breast cancer
   subtypes, along with the MCF-10A (SCSP-575) cell line, which serves as
   a model for normal breast epithelial cells, were obtained from the Cell
   Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were
   cultured in high-glucose Dulbecco's Modified Eagle's Medium
   supplemented with 10 % fetal bovine serum (JYK-FBS-301, Jin Yuan Kang
   Biotechnology) at 37 °C with 5 % CO[2]. Transient transfections of
   specific gene-targeting siRNA fragments were achieved using
   Lipofectamine 3000 (Invitrogen). The designed siRNA sequences were
   validated for their knockdown efficiency, particularly targeting
   SNHG16, to elucidate its role in breast cancer cells. Gene sequences
   are listed in [88]Table S2.

2.9. Validation of DEG expression by real-time quantitative PCR

   Total RNA was extracted from MCF-7 and MDA-MB-231 cell samples using
   the TRIzol method, and RNA concentration and purity were assessed using
   NanoDrop 2000. RNA with concentrations ranging from 200 to 2000 ng/μl
   and an OD260/OD230 ratio of 1.8–2.2 was considered suitable. RNA was
   reverse transcribed into cDNA according to kit instructions. SYBR Green
   Mix (QP031, GeneCopoeia), primers, and gene templates were combined for
   amplification, with GAPDH as the reference gene for normalization. Data
   were analyzed using the 2^-ΔΔCT method with GraphPad Prism 8.0 software
   to quantify relative gene expression levels. Gene sequences are listed
   in [89]Table S3.

2.10. Western blotting experiment

   Cells were rinsed with PBS to remove the culture medium. RIPA lysis
   buffer (20 μl) was added per culture dish to achieve complete cell
   digestion. Additionally, 2 μl of protease inhibitor and 2 μl of
   phosphatase inhibitor (Thermo Scientific) were included. Protein
   concentration was measured using the BCA method. Samples were prepared
   with 1x loading buffer, boiled, and stored at −20 °C. SDS-PAGE gel
   electrophoresis was performed to separate proteins by size. Proteins
   were then transferred to PVDF membranes. A 5 % BSA solution incubation
   blocked nonspecific binding sites. Specific monoclonal antibodies were
   used for β-actin (1:1500) and AURKA (1:1000, EM1706-71, HUABIO),
   followed by overnight incubation at 4 °C. Secondary antibodies
   (Beyotime) conjugated with horseradish peroxidase-linked rabbit
   anti-mouse IgG (1:5000) were used, followed by exposure and imaging
   using the Epizyme Omni-ECL chromogenic reagent. ImageJ software was
   employed for film scanning to quantify density and determine the
   relative expression of target genes compared to β-actin.

2.11. Scratch assay

   Cells were enzymatically digested and seeded in 6-well plates until
   reaching 80 %–90 % confluency. Using a sterile, enzyme-free 10 μl
   pipette tip, scratches were made on the cell culture surface, ensuring
   uniform width and placement under a microscope. The old medium was
   replaced with serum-free DMEM and incubated for 24 h before observation
   and photography under low magnification. Cell migration rates were
   analyzed using ImageJ software.

2.12. CCK-8 assay

   Absorbance readings measured cell viability at 24, 48, and 72 h. Cells
   were digested with trypsin, washed with PBS, and seeded uniformly in
   96-well plates (200 μl medium and 10 μl CCK-8 solution per well, NEST
   Biotechnology). After 4 h of incubation at 37 °C, absorbance at 450 nm
   was measured using a microplate reader to calculate cell proliferation
   rates.

2.13. Statistical analysis

   Sequencing data were analyzed and visualized using R 4.13. All other
   data were analyzed using SPSS 25.0, and graphs were drawn using
   GraphPad Prism 8.0.2. Quantitative data are presented as
   mean ± standard deviation. Statistical analysis employed ANOVA with a
   significance level set at α = 0.05 (two-tailed).

3. Results

3.1. Differential gene expression analysis

   We selected six breast cancer whole-genome expression datasets from the
   GEO database (lncRNA: [90]GSE65194, [91]GSE61304; miRNA: [92]GSE41922,
   [93]GSE45666; mRNA: [94]GSE29431, [95]GSE42568) based on predetermined
   criteria. To enhance the reliability of our differential gene
   selection, we intersected these results with those from the TCGA
   database. As illustrated in [96]Fig. 1A–I, upregulated genes (log[2]
   fold change (FC)>|2|, P < 0.05) are shown in red, downregulated genes
   (log[2] fold change (FC)>|2|, P < 0.05) in green, and non-significant
   genes in black. Upon intersection, we identified 112 DElncRNAs in tumor
   tissues compared to normal tissues, comprising 48 upregulated and 64
   downregulated ones; 11 DEmiRNAs with 3 upregulated and 8 downregulated;
   and 97 DEmRNAs with 47 upregulated and 50 downregulated ([97]Fig.
   1J–L).

Fig. 1.

   [98]Fig. 1
   [99]Open in a new tab

   Screening for differentially expressed genes. A, D, G: Differentially
   expressed lncRNA; B, E, H: Differentially expressed miRNA; C, F, I:
   Differentially expressed mRNA. J: Intersection of differentially
   expressed lncRNA; K: Intersection of differentially expressed miRNA; L:
   Intersection of differentially expressed mRNA.

3.2. Functional enrichment analysis of differential gene expression

   Functional enrichment analysis of DElncRNA-targeted genes revealed
   significant enrichment in protein binding, cell division, and
   cytoplasmic processes in GO analysis. KEGG pathway analysis highlighted
   enrichment in ECM receptor interaction, oocyte meiosis, Wnt/β-catenin
   signaling pathway, and cell cycle. For DEmiRNAs, GO analysis indicated
   enrichment in extracellular vesicle and membrane components, while KEGG
   pathways enriched in tyrosine metabolism, PPAR, and AMPK signaling
   pathways were noted. DEmRNA analysis showed GO enrichment in negative
   regulation of gene expression, positive regulation of cell cycle, and
   negative regulation of hematopoietic stem cells, with KEGG pathways
   enriched in Wnt/β-catenin signaling and immune system development
   ([100]Fig. 2).

Fig. 2.

   [101]Fig. 2
   [102]Open in a new tab

   Functional enrichment analysis of GO and KEGG for differential genes
   and their target genes. A: GO enrichment analysis of different
   expressed lncRNA target genes. B: KEGG enrichment analysis of different
   expressed lncRNA target genes. C: GO enrichment analysis of different
   expressed miRNA target genes. D: KEGG enrichment analysis of different
   expressed miRNA target genes. E: GO enrichment analysis of differently
   expressed mRNA. F: KEGG enrichment analysis of differently expressed
   mRNA.

3.3. Protein-protein interaction (PPI) network analysis

   Using the STRING database, we constructed PPI networks for DEmRNAs,
   filtering out isolated nodes. Separate networks were built for
   upregulated (47 nodes, 987 edges) and downregulated genes (50 nodes,
   132 edges). Further analysis using Cytoscape software plugins MCODE and
   cytoHubba identified top-scoring core proteins in each network. Key
   genes included AURKA, HMMR, FN1, RUNX2, CDK1, CCNB1, CENPA, UHRF1,
   COL1A1, NUSAP1 for upregulated genes, and FABP4, LEP, SFRP1, FMO2, OGN,
   ADH1B, GPIHBP1, ADAMTS5, TGFBR3, ZBTB16 for downregulated genes.
   Notably, AURKA was central in the PPI network, and the top 20 genes by
   score were selected for further analysis ([103]Fig. 3A–C).

Fig. 3.

   [104]Fig. 3
   [105]Open in a new tab

   PPI network and key sub-network of differentially expressed mRNAs. A:
   PPI network of up-regulated differential genes. B: PPI network of
   down-regulated differential genes. C: Key sub-networks. D: CeRNA
   regulatory network. Red represents up-regulated differentially
   expressed genes, green represents down-regulated differentially
   expressed genes. The rectangle represents mRNA, the circle represents
   miRNA, and the prism represents lncRNA. E: Complementary sequences
   between SNHG16 and hsa-let-7b-5p, AURKA and hsa-let-7b-5p.

3.4. Interaction network and ceRNA network construction

   Using TargetScan and miRanda software, we predicted target genes for
   DElncRNAs and DEmiRNAs, intersecting these results with differentially
   expressed genes. This approach yielded 13 DElncRNAs, 3 DEmiRNAs, and 3
   DEmRNAs. According to the ceRNA hypothesis, these lncRNAs should
   exhibit a negative correlation with miRNAs and a positive correlation
   with mRNAs. Constructing a breast cancer ceRNA regulatory network, we
   identified 29 interaction pairs, including 13 lncRNA-miRNA pairs, 13
   lncRNA-mRNA pairs, and 3 miRNA-mRNA pairs ([106]Fig. 3D). Each adjacent
   differential gene pair exhibited opposite expression and prognostic
   relationships. Using the StarBase database, we further evaluated the
   correlation (R) strength between differential genes, focusing on the
   most robust ceRNA subnetwork (SNHG16-hsa-let-7b-5p-AURKA) for
   subsequent experimental validation ([107]Table S4).

3.5. Differential gene interactions in ceRNA regulatory networks

   Using RNA chip technology, we identified complementary binding sites
   between lncRNA SNHG16 and miRNA hsa-let-7b-5p ([108]Fig. 3E). This
   suggests SNHG16's potential involvement in the ceRNA regulation of
   hsa-let-7b-5p in breast cancer. Additionally, hsa-let-7b-5p was found
   to interact with mRNA AURKA, indicating its role in regulating mRNA
   expression via ceRNA mechanisms.

3.6. Expression of differential genes in ceRNA networks in breast cancer

   To validate ceRNA differential gene expression in breast cancer, we
   evaluated the prognostic value of SNHG16 and AURKA and hsa-let-7b-5p
   using the Kaplan-Meier database, and we assessed their expression using
   the UALCAN database ([109]Fig. 4A–C). To compare the expression of
   AURKA, we selected breast tissue sections from normal individuals
   (CAB001454, ID: 278) and breast cancer patients (HPA002636, ID: 2174)
   registered in the Proteinatlas database ([110]Fig. 4D). All results
   showed elevated SNHG16 and AURKA expression, and decreased
   hsa-let-7b-5p expression in breast cancer, correlating with prognosis
   and supporting ceRNA network mechanisms.

Fig. 4.

   [111]Fig. 4
   [112]Open in a new tab

   Differential gene expression. A: Survival analysis of AURKA in breast
   cancer prognosis. B: Survival analysis of SNHG16 in breast cancer
   prognosis. C: Survival analysis of hsa-let-7b-5p in breast cancer
   prognosis. D: AURKA expression in normal and breast cancer tissue
   sections. E: Expression of differential genes in breast cancer cells
   and normal breast epithelial cells (n = 4). ∗P < 0.05, ∗∗P < 0.01,
   ∗∗∗P < 0.001.

3.7. qRT-PCR experimental results

3.7.1. qRT-PCR validation of differential gene expression in breast cancer
cells

   We performed qRT-PCR to detect SNHG16, hsa-let-7b-5p, and AURKA
   expression in two breast cancer cell lines (MCF-7 and MDA-MB-231) and
   normal breast epithelial cells (MCF-10A). As shown in [113]Fig. 4E,
   SNHG16 and AURKA were significantly upregulated in breast cancer cells,
   while hsa-let-7b-5p was downregulated. SNHG16 showed more pronounced
   upregulation in MDA-MB-231 cells, and hsa-let-7b-5p downregulation was
   more evident in these cells, consistent with our predictions.

3.7.2. Knockdown of SNHG16 and qRT-PCR validation of gene knockdown
efficiency

   To further investigate the role of lncRNA in ceRNA regulation, we
   designed si-SNHG16 interference fragments to knock down SNHG16
   expression in cells. Fluorescence microscopy confirmed transfection
   efficiency above 90 %, indicating successful transfection for
   subsequent experiments. qRT-PCR revealed a significant reduction in
   SNHG16 expression post-transfection in MCF-7 cells, accompanied by
   decreased AURKA expression and increased hsa-let-7b-5p expression.
   Similar trends were observed in MDA-MB-231 cells, with notably
   increased hsa-let-7b-5p expression ([114]Fig. 5A–C).

Fig. 5.

   [115]Fig. 5
   [116]Open in a new tab

   Effect of knockdown SNHG16 expression on ceRNA. A: Efficiency of
   transfection of siRNA under fluorescence microscopy. B: qRT-PCR
   detection of transfection efficiency of si-SNHG16 on MCF-7 cells
   (n = 3). C: qRT-PCR detection of transfection efficiency of si-SNHG16
   on MDA-MB-231 cells (n = 3). D, E: Effect of knockdown of SNHG16 on
   proliferation ability of MCF-7 and MDA-MB-231 in breast cancer cells
   (n = 4). ∗P < 0.05, ∗∗P < 0.01.

3.8. Impact of SNHG16 knockdown on proliferation and migration of breast
cancer cells

3.8.1. Inhibition of breast cancer cell proliferation by SNHG16 knockdown

   CCK-8 assays demonstrated that knocking down SNHG16 inhibited breast
   cancer cell proliferation ([117]Fig. 5D–E). The most significant
   inhibition was observed at 48 h, indicating SNHG16's role in promoting
   breast cancer proliferation in both MCF-7 and MDA-MB-231 cells.

3.8.2. Inhibition of breast cancer cell migration by SNHG16 knockdown

   Scratch assays showed that SNHG16 knockdown suppressed breast cancer
   cell migration ([118]Fig. 6A–B). This further supports SNHG16's role in
   promoting breast cancer cell migration.

Fig. 6.

   [119]Fig. 6
   [120]Open in a new tab

   Effect of lncRNA SNHG16 on cell migration ability. A: Effect of
   knockdown of SNHG16 expression on breast cancer cell migration. B:
   Inhibition rate of breast cancer cell migration by knockdown of SNHG16
   expression (n = 3). C. AURKA protein expression after knockdown of
   SNHG16 in both cell lines (n = 3). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001,
   ns represents p-value is not significant.

3.9. Verification of AURKA protein expression by western blotting

   To validate the expression of AURKA protein following the knockdown of
   lncRNA SNHG16 in MCF-7 and MDA-MB-231 cells, Western blot experiments
   were conducted. The results indicated a significant decrease in AURKA
   protein expression upon SNHG16 knockdown in both cell lines ([121]Fig.
   6C). These findings corroborate with results from the GEPIA database
   and qRT-PCR analysis, confirming SNHG16's role in modulating AURKA
   protein expression through ceRNA mechanisms. This regulatory pathway is
   pivotal in influencing breast cancer cell proliferation and migration,
   thereby impacting disease progression and prognosis.

4. Discussion

   Breast cancer is the most prevalent tumor among women globally and
   ranks as the second most common cancer worldwide. Incidence and
   mortality rates are increasing annually [[122]1]. The disease often
   begins subtly, making early detection challenging, with 90 % of breast
   cancer patients succumbing to metastasis [[123]31]. In this study, we
   conducted gene screening using TCGA and GEO databases to identify
   differentially expressed genes in breast cancer. Functional enrichment
   analyses of these genes and their targets revealed significant
   involvement in cell division-related activities, underscoring the high
   metastatic potential of breast cancer.

   SNHG16 was initially detected to be highly expressed in invasive
   neuroblastomas [[124]32]. Recent experiments and sequencing results
   have consistently shown elevated SNHG16 expression in various cancers
   such as liver, colorectal, and breast cancers, aligning with our
   expectations [[125][33], [126][34], [127][35]]. Studies indicate that
   high SNHG16 expression contributes to breast cancer progression through
   multiple mechanisms, including promoting cell proliferation,
   metastasis, angiogenesis, and chemoresistance [[128][36], [129][37],
   [130][38], [131][39], [132][40]]. In breast cancer, SNHG16 was newly
   found to act as a novel biomarker through ceRNA interactions
   [[133][41], [134][42], [135][43], [136][44]]. AURKA positively
   regulates the G2 to M transition in the cell cycle and spindle-related
   activities during early mitosis [[137]44]. Its interactions with
   substrates like histone H3 and proteins involved in centrosome
   separation and spindle formation underscore its essential role in
   mitotic progression, highlighting its role in breast cancer progression
   [[138][45], [139][46], [140][47]]. For example, AURKA collaborates with
   HMMR in spindle assembly during mitosis [[141][48], [142][49],
   [143][50]], while CDK1 stabilizes AURKA expression, enhancing its role
   in cell cycle regulation [[144]51]. The reliability of these core
   differential genes and subnetworks was validated through multiple
   experimental studies, emphasizing their critical roles in cancer
   progression. However, comprehensive studies integrating large-scale
   gene datasets and experimental validation are needed to elucidate their
   roles in breast cancer progression. Integration of predicted target
   genes yielded a final set of differentially expressed genes, forming a
   ceRNA regulatory network. We identified SNHG16-hsa-let-7b-5p-AURKA as
   the most robustly correlated ceRNA network in breast cancer. Previous
   studies have validated SNHG16's role as an upstream regulator of
   hsa-let-7b-5p in hepatocellular carcinoma [[145]52]. In contrast,
   AURKA's downstream influence on hsa-let-7b-5p has been observed in
   neuroglial tumors [[146]53], affirming the regulatory presence of this
   ceRNA in cancer control. However, the potential role of this regulatory
   axis in breast cancer has not been demonstrated.

   Prognostic analysis of DEGs within the ceRNA network in breast cancer
   revealed high expression levels of SNHG16 and AURKA correlating with
   poor prognosis, whereas low hsa-let-7b-5p expression indicated
   unfavorable outcomes. There have been experimental validations have
   confirmed AURKA's overexpression in ER-positive and triple-negative
   breast cancers, which is similar to our experimental results
   [[147]54,[148]55]. In contrast, low hsa-let-7b-5p levels correlate with
   poorer prognosis in breast cancer, possibly due to its negative
   correlation with SMUG1 expression, impacting survival rates in estrogen
   receptor-positive breast cancer [[149]56]. SNHG16 is implicated in
   several pathways to modulate tumorigenesis; for instance,
   downregulating SNHG16 in esophageal cancer inhibits the Wnt/β-catenin
   pathway, thereby suppressing cell proliferation and invasion [[150]57].
   Similarly, reduced SNHG16 expression in bladder cancer regulates
   downstream proteins (Wnt1, c-myc, cyclin-D1, and catenin) through the
   Wnt/β-catenin pathway to inhibit T24 cell proliferation and migration
   [[151]58]. SNHG16 also enhances TGF-β1/SMAD5 pathway activity via the
   miR-16-5P and SMAD5 ceRNA axis, promoting breast cancer development
   [[152]59].

   We experimentally confirmed the impact of SNHG16 downregulation on cell
   proliferation and migration. Our findings revealed varying inhibition
   in MCF-7 and MDA-MB-231 cell lines, consistent with previous studies
   [[153]59,[154]60]. Specifically, MDA-MB-231 cells showed pronounced
   inhibition of proliferation, while MCF-7 cells exhibited greater
   suppression of migration, likely due to differences in cellular
   morphology and growth rates, necessitating further experimental
   validation. However, our study did not analyze the expression of
   differential genes across all breast cancer subtypes or the effect on
   breast cell proliferation after the knockout of other genes on this
   axis, warranting further exploration.

5. Conclusion

   In conclusion, we successfully constructed a prognostically relevant
   ceRNA regulatory network in breast cancer through differential gene
   screening. Experimental validation confirmed SNHG16-hsa-let-7b-5p-AURKA
   in breast cancer cells and affirmed that SNHG16 promotes cell
   proliferation and migration.

CRediT authorship contribution statement

   Zipeng Qiao: Writing – original draft, Visualization, Validation,
   Software, Methodology, Conceptualization. Yongjun Tang: Validation,
   Software, Methodology, Data curation, Conceptualization. Shengle Li:
   Visualization, Validation, Software. Qiufeng Lao: Software,
   Methodology, Investigation, Conceptualization. Yu Xing: Visualization,
   Software, Conceptualization. Qingquan Zhang: Methodology,
   Investigation, Formal analysis. Songying Pan: Validation, Methodology.
   Chaoyong Bei: Visualization, Resources, Investigation. Weiyi Pang:
   Writing – review & editing, Project administration, Funding
   acquisition, Conceptualization. Hui Liu: Writing – review & editing,
   Supervision, Project administration, Conceptualization.

Declaration of competing interest

   The authors declare that they have no known competing financial
   interests or personal relationships that could have appeared to
   influence the work reported in this paper.

Acknowledgments