Synergistic inhibition of colorectal cancer progression by silencing Aurora A and the targeting protein for Xklp2

Abstract

BACKGROUND

Unraveling the pathogenesis of colorectal cancer (CRC) can aid in developing prevention and treatment strategies. Aurora kinase A (AURKA) is a key participant in mitotic control and interacts with its co-activator, the targeting protein for Xklp2 (TPX2) microtubule nucleation factor. AURKA is associated with poor clinical outcomes and high risks of CRC recurrence. AURKA/TPX2 co-overexpression in cancer may contribute to tumorigenesis. Despite its pivotal role in CRC development and progression, the action mechanism of AURKA remains unclear. Further research is needed to explore the complex interplay between AURKA and TPX2 and to develop effective targeted treatments for patients with CRC.

AIM

To compare effects of AURKA and TPX2 and their combined knockdown on CRC cells.

METHODS

We evaluated three CRC gene datasets about CRC (GSE32323, GSE25071, and GSE21510). Potential hub genes associated with CRC onset were identified using the Venn, search tool for the retrieval of interacting genes, and KOBAS platforms, with AURKA and TPX2 emerging as significant factors. Subsequently, cell models with knockdown of AURKA, TPX2, or both were constructed using SW480 and LOVO cells. Quantitative real-time polymerase chain reaction, western blotting, cell counting kit-8, cell cloning assays, flow cytometry, and Transwell assays were used.

RESULTS

Forty-three highly expressed genes and 39 poorly expressed genes overlapped in cancer tissues compared to controls from three datasets. In the protein-protein interaction network of highly expressed genes, AURKA was one of key genes. Its combined score with TPX2 was 0.999, and their co-expression score was 0.846. In CRC cells, knockdown of AURKA, TPX2, or both reduced cell viability and colony number, while blocking G0/G1 phase and enhancing cell apoptosis. Additionally, they were weakened cell proliferation and migration abilities. Furthermore, the expression levels of B-cell lymphoma-2-Associated X, caspase 3, and tumor protein P53, and E-cadherin increased with a decrease in B-cell lymphoma-2, N-cadherin, and vimentin proteins. These effects were amplified when both AURKA and TPX2 were concurrently downregulated.

CONCLUSION

Combined knockdown of AURKA and TPX2 was effective in suppressing the malignant phenotype in CRC. Co-inhibition of gene expression is a potential developmental direction for CRC treatment.

Key Words: Aurora kinase A; Targeting protein for Xklp2; Microtubule nucleation factor; Colorectal cancer; Proliferation; Migration; Invasion

Core Tip: In this study, we evaluated three gene datasets and identified 13 crucial genes. Results from SW480 and LOVO cells showed that Aurora kinase A (AURKA) and the targeting protein for Xklp2 (TPX2) blocked cell cycle progression and promoted cell apoptosis of colorectal cancer (CRC). Knockdown of either AURKA, TPX2, or both reduced the invasion and migration ability of CRC cells while inhibiting epithelial-mesenchymal transition. Co-knockdown of AURKA and TPX2 enhanced the inhibition of CRC cells. Combined knockdown was more effective in suppressing the malignant phenotype in CRC, providing ideas and basic data for CRC treatment.

INTRODUCTION

Globally, colorectal cancer (CRC) is responsible for the top three number of new cancer cases and cancer-related deaths[1], and is also the top three most common cancer in both males and females. The rectum and sigmoid colon are the most common locations in CRC[2]. While the overall occurrence in individuals over 50 years old stabilizes or declines with improved health standards, younger cohorts still face growing risks[3]. Several factors increase the risk of early-onset CRC[4], including genetic predisposition, poor diet, lack of exercise, smoking, obesity, and diabetes. The precise mechanism of CRC has yet to be fully studied, which limits the development of treatment strategies. Therefore, unraveling the pathogenesis of CRC can help develop new prevention and treatment strategies, and reduce the social burden.

The Aurora kinase A (AURKA) is a key participant in mitotic control[5]. AURKA activity can be regulated by the self-phosphorylation of Thr288 and its interaction with its coactivator, the targeting protein for Xklp2 (TPX2) microtubule nucleation factor[6]. AURKA is implicated in regulating cellular processes, such as proliferation and angiogenesis[7]. In patients with CRC, AURKA is frequently amplified and overexpressed[8], which has been linked to poor clinical outcomes and a higher risk of recurrence. However, a recent study suggested that high expression of AURKA was positively correlated with overall survival in patients with stage II CRC[9]. Researchers have discovered that AURKA overexpression alone does not cause its interphase nucleus accumulation; instead, high accumulation requires AURKA and TPX2 co-overexpression, suggesting that co-overexpression may contribute to tumorigenesis[10]. Despite its pivotal role in CRC development and progression, further research is still needed to explore the complex interplay between AURKA and TPX2 and to develop effective targeted treatments for patients with CRC. Therefore, we compared the effects of single and combined knockdowns on CRC malignant phenotype through in vitro experiments.

MATERIALS AND METHODS

Dataset analysis

Collecting data: National center of biotechnology information gene expression omnibus (GEO) is a professional genetic database that provides gene datasets for cancer and other diseases. Using CRC as the keyword, we collected three relevant datasets from the GEO database (GSE32323, GSE25071, and GSE21510) (https://www.ncbi.nlm.nih.gov/geo/) (their information is provided in Table 1).

Table 1 Basic information on the gene expression omnibus dataset included in the study.

GEO	Group	n	Stage	Tissue	Gene number
GSE21510	CN	12	0-2	Normal	715
	CRC	8	1-2b	Cancer
GSE25071	CN	3	0	Colon	788
	CRC	7	I-II	Primary tumor
GSE32323	CN	9	1-2	Normal	446
	CRC	9	1-2b	Cancer

GEO: Gene expression omnibus; CN: Control; CRC: Colorectal cancer.

Analyzing data: Screening genes were based on |Log₂ fold change (FC)|> 2 and adjusted P value < 0.05. Based on normal tissues, low- and high-expression groups were analyzed separately to perform the intersection of three datasets. The potential target genes were screened using network pharmacological analysis. Protein-protein interaction (PPI) networks and Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analysis were performed using Venn, search tool for the retrieval of interacting genes (STRING), and KOBAS.

Visualization: Visualization tools were used for both the data analysis and presentation. OmicShare tools were used to generate Venn plots (https://www.omicshare.com/tools/home/report/reportvenn.html). STRING (version 12.0)[11] was a website to integrate all publicly available PPI information sources. Cytoscape (version 3.9.1)[12] was used to analyze the potential target genes. KOBAS (version 3.0)[13] was a website that can perform and provide KEGG pathway enrichment information. Subsequently, we used the SRplot bubble chart tool (http://www.bioinformatics.com.cn/srplot) to visualize the KEGG enrichment results.

Expression analysis of target genes: We obtained gene expression levels and overall survival curves using gene expression profiling interactive analysis (GEPIA)[14]. Photographs showing AURKA protein expression in human colonic adenocarcinoma (COAD) tissues and normal colon samples were obtained from the human protein atlas (https://www.proteinatlas.org/). The website for obtaining normal colon tissue photos is https://www.proteinatlas.org/ENS G00000087586-AURKA/tissue/colonimg; the website for obtaining COAD tissue photos is: (https://www.proteinatlas.org/ENSG00000087586-AURKA/pathology/colorectal+cancerimg).

Cell transfection and grouping

SW480 (iCell-h204; iCell Bioscience, China), HCT116 (iCell-h071; iCell Bioscience), and LOVO (iCell-h126; iCell Bioscience, China) are human CRC cell lines; the NCM460 (iCell-h373; iCell Bioscience) is a human normal colon epithelial cell line. They were mycoplasma-free and were identified by short tandem repeat analysis. Cell culture and storage were performed in a basic culture medium (SH30022.02; HyClone, United States) and 10% fetal bovine serum (FBS) (SH30084.03; HyClone, Australia). All the cell lines kept at 37 °C and 5% carbon dioxide, a humidified incubator.

Lipo6000 transfection reagent (C0526-1.5 mL, Beyotime, China), plasmids carried short hairpin (sh) RNA (shRNA) against AURKA or TPX2 mRNA, and blank vector plasmid (Genechem Co., Ltd., China) was used for transfection. Briefly, cells were incubated with 125 μL premixed liquid containing Lipo6000, Dulbecco’s modified eagle medium, and 2.5 mg shRNA for 25 minutes. The dosage was based on the Lipo6000 instructions. After 48 hours of transfection, quantitative real-time polymerase chain reaction (qPCR), as well as western blotting (WB) were used for the quantitative analysis of transfection efficiency. SW480 and LOVO cells were divided into four groups: Negative control (NC) of shRNA (sh-NC), sh-AURKA, sh-TPX2, and sh-AURKA + sh-TPX2.

Quantitative relative transcript levels

Total RNA was extracted from SW480 and LOVO cells using accurate biology reagents (AG21024; China). The mixed liquid of cells and lysate was added to 70% ethanol and used to collect total RNA through multiple centrifugations and elution according to instructions of RNA instructions. Subsequently, the RNA was subjected to reverse transcription and qPCR using a HiFiScript cDNA synthetic kit (CW2569M, CWBIO, China) and a SYBR Green qPCR kit (11201ES03, Yesaen, China) according to the kits’ instructions. All the reactions were performed using a LightCycler 96 instrument (Roche, Germany). Human beta-actin gene served as a housekeeping gene. Primer sequences are listed in Table 2 and the 2^-ΔΔCt method were used for qPCR.

Table 2 Sequence information of the primers for quantitative real time-polymerase chain reaction.

Gene	Forward primer	Reverse primer
Human AURKA	5’-TGGGTGGTCAGTACATGCTC-3’	5’-TGCATCCGACCTTCAATCATTTC-3’
Human TPX2	5’-ACTTCCGCACAGATGAGCG-3’	5’-GGATGCTTTCGTAGTTCAGATGT-3’
Human β-actin	5’-GATGACCCAGATCATGTTTGAG-3’	5’-TAATGTCACGCACGATTTCC-3’

AURKA: Aurora kinase A; TPX2: Targeting protein for Xklp2.

Quantitative protein levels

Total protein samples from SW480 and LOVO cells were collected using a lysis solution (89901, Thermo, United States). The samples and loading buffer (P10015, Beyotime, China) were mixed at a ratio of 1:5 and boiled for 5 minutes. Subsequently, 20 μL samples were separated and transferred to activated PVDF membranes (10600023, GE Healthcare Life, United States) using electrophoresis. After blocking, the membranes were incubated with antibodies. The antibodies used are listed in Table 3. Finally, the membranes were visualized and analyzed using a gel imaging instrument (610020-9Q, Qing Xiang, China) and ImageJ software (NIH, United States).

Table 3 Antibody information.

Antibody	Manufacturer	No.	Dilution ratio
AURKA antibody	Affinity	BF0123	1:2000
TPX2 antibody	Abcam	ab264124	1:5000
Cyclin B1 antibody	Affinity	AF6168	1:2000
CDK1 antibody	Abcam	ab133327	1:15000
CDK2 antibody	Affinity	AF6237	1:2000
Bax antibody	Abcam	ab182733	1:2000
Bcl-2 antibody	Affinity	AF6139	1:2000
Caspase-3 antibody	Proteintech	19677-1-AP	1:10000
p53 antibody	Affinity	AF0879	1:2000
N-cadherin antibody	Affinity	AF4039	1:2000
E-cadherin antibody	Proteintech	20874-1-AP	1:5000
Vimentin antibody	Zenbio	R22775	1:1000
p-AURKA antibody	Affinity	AF3011	1:2000
β-actin antibody	Proteintech	81115-1-RR	1:20000
Anti-rabbit IgG, HRP-linked antibody	CST	7074	1:10000

AURKA: Aurora kinase A; TPX2: Targeting protein for Xklp2; Bcl-2: B-cell lymphoma-2; Bax: B-cell lymphoma-2 associated X protein; p53: Tumor protein P53; p-AURKA: Phosphorylated Aurora kinase A; HRP: Horseradish peroxidase; CST: Cell signaling technology.

Cell viability and proliferation evaluation

A cell counting kit-8 (CCK-8) and cell cloning assays were used to evaluate cell proliferation. The CCK-8 kit (C0037; Beyotime, China) was used according to the manufacturer’s instructions. Briefly, transfected cells (1 × 10³ cells/well) were cultured with complete mediums for 12 hours, 24 hours, and 48 hours. Fresh culture medium containing 10 μL CCK-8 reagent replaced the complete medium incubating for 4 hours. The optical density (OD) at 450 nm was directly proportional to the cell viability and was detected using a CMaxPlus microplate reader (Molecular Devices, United States). Transfected cells (400 cells/well, 24-well plate) were cultured for 10 days and then fixed. Formed colonies were stained with 0.1% crystal violet (548-62-9; Qiangshun, China) for 2 minutes. The number of colonies (over 20 cells)[15] was counted using an optical microscope (AE2000; Motic, China).

Detection of cell cycle and apoptosis

Flow cytometry (FCM) was used to evaluate the cell cycle progression and the rate of apoptosis. After 48 hours of transfection, cells were collected using trypsin and centrifuged. After washing with phosphate buffered saline, the cells were fixed with 75% ethanol at -20 °C for 2 hours. Then, cells were centrifugated at 200 g for 10 minutes and were incubated with 0.25% Triton X-100 at 5 °C for 4 minutes. After centrifugation, the cells (1 × 10⁷ cells/mL) were stained with propidium iodide/RNase staining buffer (550825, BD, United States) in the dark for 30 minutes. After the cells were washed, centrifuged, and resuspended, the cell cycle was analyzed using a NovoCyte flow cytometer (Agilent, United States). For apoptosis assessment, an Apoptosis detection kit (556547, BD Biosciences) was used. After dealing with working reagents, each tube with 100 μL cells (1 × 10⁶ cells/mL) was added 400 μL binding buffer to analyze using a flow cytometer (NovoCyte, Agilent, United States).

Evaluation of cell migration and invasion

Transwell assays were performed using a Transwell chamber (8.0 μm pore, Corning, United States) with Matrigel (356234, BD, United States) or not for invasion or migration assay. Cells (5 × 10⁴) were seeded in the upper chamber with basal medium and 1% FBS. The lower chamber was supplemented with basal medium and 10% FBS. Whether the migration assay or invasion assay, cells were cultured for 24 hours, and then were blocked and stained with 4% paraformaldehyde and 0.1% crystal violet. The cells were counted using an optical microscope (ICX41, Sunny, China).

Statistical analysis

We used statistical product and service solutions (version 20.0; IBM Corp., Armonk, NY, United States) and GraphPad Prism 9 (GraphPad software, San Diego, CA, United States) for statistical analysis. All data were presented as mean ± SD. Measurements from multiple datasets were analyzed via one-way analysis of variance and the Tukey post hoc test. A significance level threshold of less than 0.05 denoted a statistically relevant finding.

RESULTS

AURKA and TPX2 are potential key genes for regulating the malignant phenotype of CRC

After analysis of the CRC gene chip data (GSE21510, GSE25071, and GSE32323), we obtained three gene lists for comparison between normal tissues and CRC tissues. There were 715, 788, and 466 genes in GSE21510, GSE25071, and GSE32323 datasets, respectively (Table 1). Based on the screening criteria |Log₂ (FC)| > 2 and an adjusted P value < 0.05, we identified genes whose expression levels showed significant differences between normal and CRC tissues. GSE21510: Highly expressed genes (246) and downregulated genes (368); GSE25071: Highly expressed genes (360) and downregulated genes (425); GSE32323: Highly expressed genes (153) and downregulated genes (243). Subsequently, Venn diagrams were used to analyze highly/lowly expressed genomic intersections separately. There are 43 highly expressed genes and 39 poorly expressed genes overlapped in the three datasets (Figure 1A and B). Highly or weakly expressed genes were analyzed using STRING and two PPI networks were obtained (Supplementary Figure 1). As the PPI network of the highly expressed gene interaction network was more complex, we chose the PPI network of highly expressed genes for Cytoscape analysis (Figure 1C). The average number of neighboring nodes was nine. Thirteen key genes (degree > 9) were identified: CCNB1, CDK1, CDKN3, UBE2C, ANLN, AURKA, CENPA, CKS2, TPX2, CENPN, TRIP13, CKAP2, and CXCL10 (Table 4). In addition, AURKA and TPX2 showed strong interactions with each other (Table 5). STRING analysis showed that the combined score of AURKA and TPX2 was 0.999, and their co-expression score was 0.846.

Figure 1 Protein-protein interaction network of highly expressed genes in colorectal cancer tissues. A: A Venn diagram showed the number of relative highly expressed genes from the GSE32323, GSE25071, and GSE21510 datasets. These gene expressions were increased in colorectal cancer (CRC) tissues compared to normal tissues and an intersection of 43 genes was identified across all three datasets; B: A Venn diagram showed the number of genes with relatively low expression from GSE32323, GSE25071, and GSE21510. These gene expressions were decreased in CRC tissues compared to normal tissues and an intersection of 39 genes was identified across all three datasets. All genes were screened for |Log₂ (fold change)|> 2 and adjusted P value < 0.05; C: The topology of the protein-protein interaction network was analyzed by Cytoscape and the network was established based on 43 genes from an intersection of relative overexpressed genes from the GSE32323, GSE25071, and GSE21510 datasets. The greater the adjacent node count, the darker the green; the thicker the edges, the stronger the interaction force.

Table 4 Potential target proteins and their degrees in the relative overexpressed gene set analyzed by Cytoscape.

No.	Node	Degrees
1	CCNB1	14
2	CDK1	14
3	CDKN3	13
4	UBE2C	13
5	ANLN	12
6	AURKA	12
7	CENPA	12
8	CKS2	12
9	TPX2	12
10	CENPN	11
11	TRIP13	11
12	CKAP2	10
13	CXCL10	10

Table 5 The search tool for the retrieval of interacting genes analysis results of protein-protein interaction forces (TOP 7, score = 0.999).

Node 1	Node 2	Co-expression	Experimentally determined interaction	Database annotated	Combined score
CCNB1	CDK1	0.895	0.999	0.9	0.999
AURKA	TPX2	0.846	0.988	0.9	0.999
CXCL10	CXCL11	0.824	0.994	0.9	0.999
CDK1	CKS2	0.689	0.992	0	0.999
CCNB1	CKS2	0.688	0.978	0	0.999
CENPA	CENPN	0.407	0.955	0.5	0.999
CXCL11	CXCL2	0.107	0.994	0.5	0.999

Additionally, AURKA takes part in the progesterone-mediated oocyte maturation (P = 0.00129) and oocyte meiosis (P = 0.013128) pathways (Figure 2). Furthermore, we retrieved AURKA and TPX2 from GEPIA and found that they were highly expressed in COAD tissues (Figure 3A and B). Scientists have found a significant correlation between AURKA expression and a low survival rate in CRC liver metastasis (hazard ratio = 1.55, P < 0.01)[16]. Furthermore, AURKA expression was observed in COAD clinical samples (Figure 3C). The human protein atlas showed that in colon gland cells of normal tissues, AURKA was expressed at low levels (positive rate < 25%), whereas it was highly expressed in CRC (Figure 3C). Scientific research confirms the coexistence of AURKA and TPX2 in CRC and that they participate in the occurrence of CRC[17]. Thus, our study focused on these proteins in the malignant phenotype of CRC cells.

Figure 2 Bubble plot of Kyoto encyclopedia of genes and genomes pathway enrichment analysis results. All pathways had a corrected P value < 0.05. The KOBAS 3.0 (http://bioinfo.org/kobas/genelist/) was used for Kyoto encyclopedia of genes and genomes pathway enrichment analysis of 43 genes with relatively high expression and 39 genes with relatively low expression. The order of the pathways was based on -log₁₀ (P value). The red rectangle marks the pathways in which Aurora A participated. KEGG: Kyoto encyclopedia of genes and genomes; IL: Interleukin; TNF: Tumor necrosis factor; TGF: Transforming growth factor; PPAR: Peroxisome proliferators-activated receptor.

Figure 3 Aurora A and targeting protein for Xklp2 are high expressed in human colonic adenocarcinoma tissue. A: The Aurora kinase A (AURKA) expression in colonic adenocarcinoma (COAD) tissue was increased compared to normal tissues; B: The targeting protein for Xklp2 expression in COAD tissue was increased compared to normal tissues. The expression level analysis results are from gene expression profiling interactive analysis; C: Expression of the AURKA protein in human COAD tissue sample or normal colon. The normal colon tissue photos from (https://www.proteinatlas.org/ENSG00000087586-AURKA/tissue/colonimg); the COAD tissue photos from the human protein atlas (https://www.proteinatlas.org/ENSG00000087586-AURKA/pathology/colorectal+cancerimg). ^aP < 0.05. AURKA: Aurora kinase A; TPX2: Targeting protein for Xklp2; COAD: Colonic adenocarcinoma; CRC: Colorectal cancer; T: Tumors; N: Normal.

AURKA and TPX2 knockdown cell lines were constructed successfully

Transcript levels were measured using qPCR. HCT116, SW480, and LOVO are CRC cells with relatively high expression levels of AURKA and TPX2 compared to NCM460, a human colon epithelial cell line (Figure 4A). Among them, AURKA and TPX2 showed the highest expression levels in SW480 and LOVO cells (Figure 4A). Therefore, we used SW480 and LOVO cell lines for AURKA and TPX2 knockdown models. Their transcript levels were determined by qPCR and protein expression was measured by WB. The transcript levels of AURKA and TPX2 in SW480 cells with sh-AURKA3 treatment decreased to 33.3% and 48.0% (P < 0.01); they in LOVO cells with sh-TPX23 treatment decreased to 28.9% and 28.3%, respectively (P < 0.01) (Figure 4B and C). In AURKA and TPX2 knockdown cell models, the decreasing trend of AURKA and TPX2 proteins was consistent with their transcription levels (P < 0.01) (Figure 4D and E).

Figure 4 Knockdowns of Aurora A and targeting protein for Xklp2 inhibited viability of colorectal cancer cell. A: Relative expression of Aurora kinase A (AURKA) and targeting protein for Xklp2 (TPX2) in NCM460, SW480, HCT116, and LOVO cells; B: After short hairpin RNA of AURKA (sh-AURKA) 1, 2, and 3 plasmids transfection, AURKA mRNA expression was decreased in SW480 and LOVO cells; C: TPX2 mRNA expression was decreased in SW480 and LOVO cells with transfected sh-TPX21, 2, and 3 plasmids. All gene transcript levels were measured by quantitative real time-polymerase chain reaction; D: AURKA; E: TPX2 protein expression levels were decreased in SW480 and LOVO cells with sh-AURKA/TPX21, 2, and 3 plasmids transfection. Western blotting was used to detect relative protein expression. Cell counting kit-8 assay was used to detect cell viability in F: SW480; G: LOVO cells with AURKA and TPX2 knockdown. After 12-, 24-, or 48-hour treatment with transfection, cell viability was inhibited in SW480 and LOVO cells. ¹P < 0.01 vs NCM460 group. ²P < 0.05 vs negative control of short hairpin RNA (sh-NC) group. ³P < 0.01 vs sh-NC group. ⁴P < 0.05 vs short hairpin RNA of AURKA (sh-AURKA) group. ⁵P < 0.01 vs sh-AURKA group. ⁶P < 0.01 vs short hairpin-TPX2 group. qPCR: Quantitative real time-polymerase chain reaction; AURKA: Aurora kinase A; TPX2: Targeting protein for Xklp2; sh-NC: Negative control of short hairpin RNA; sh-AURKA: Short hairpin RNA of Aurora kinase A; sh-TPX2: Short hairpin RNA of targeting protein for Xklp2; OD: Optical density; WB: Western blotting.

AURKA and TPX2 knockdown inhibited proliferation and disrupted the cell cycle in CRC

Subsequently, the proliferation capacity was analyzed using CCK-8 and cell cloning assays. We also used WB to quantify biomarker levels in the cell cycle. The OD value is directly proportional to cell viability. We found that after treatment with sh-AURKA, sh-TPX2, or a combination of both for 12 hours, 24 hours, or 48 hours, the cell viability was significantly inhibited (P < 0.01), whereas no statistical differences emerged when comparing the sh-NC group to the control (Figure 4F and G). Specifically, AURKA and TPX2 co-knockdown exhibited the strongest impact on cell viability (P < 0.01) (Figure 4F and G) and significantly inhibited cell viability compared with sh-AURKA treatment (P < 0.05) (Figure 4F and G). Cell cloning assays validated this observation and revealed a marked decrease in colony number in the sh-AURKA, sh-TPX2, and sh-AURKA + sh-TPX2 groups compared to the sh-NC group (P < 0.05) (Figure 5A). Furthermore, FCM revealed a significant decrease G0/G1 cell percentage and an increase in G2/M percentage among SW480 and LOVO cells treated with sh-AURKA, sh-TPX2, or their combination, while G2/M cell percentage was significantly increased (P < 0.01) (Figure 5B and C), paired with significantly reduced expression of key biomarkers such as cyclin B1, CDK1, and CDK2 (P < 0.05) (Figure 5D and E).

Figure 5 Knockdowns of Aurora A and targeting protein for Xklp2 inhibited colorectal cancer cell proliferation ability and decreased cell percentage in the G0/G1 phase. A: The cell cloning assay was used to determine cell proliferation ability. Knockdown or co-knockdown of Aurora kinase A and targeting protein for Xklp2 (TPX2) inhibited the number of cell colonies; B: Cell cycle in SW480 detected by flow cytometry; C: Cell cycle in LOVO cells was detected by flow cytometry. Knockdown or co-knockdown AURKA and TPX2 decreased cell percentage in the G0/G1 phase; D: Western blotting was used to detect the expression levels of cell cycle-related biomarkers. Cyclin B1, CDK1, and CDK2 expressions were decreased in SW480 and LOVO cells with AURKA and TPX2 co-knockdown; E: The representative bands of western blotting. ¹P < 0.05 vs negative control of short hairpin RNA (sh-NC) group. ²P < 0.01 vs sh-NC group. ³P < 0.05 vs short hairpin RNA of AURKA (sh-AURKA) group. ⁴P < 0.01 vs sh-AURKA group. ⁵P < 0.05 vs short hairpin-TPX2 (sh-TPX2) group. ⁶P < 0.01 vs sh-TPX2 group. AURKA: Aurora kinase A; TPX2: Targeting protein for Xklp2; sh-NC: Negative control of short hairpin RNA; sh-AURKA: Short hairpin RNA of Aurora kinase A; sh-TPX2: Short hairpin RNA of targeting protein for Xklp2.

AURKA and TPX2 knockdown promoted cell apoptosis and inhibited cell migration and invasion

After the downregulation of AURKA, TPX2, or their combination in SW480 and LOVO cells, FCM revealed significantly enhanced apoptosis (P < 0.01) (Figure 6A). In particular, TPX2 knockdown was more potent at inducing apoptosis (P < 0.05), and TPX2 and AURKA co-knockdown had the best effect (P < 0.05) (Figure 6A). WB revealed that B-cell lymphoma-2 (Bcl-2)-associated X protein (Bax), caspase 3, and tumor protein P53 (p53) levels were significantly increased following silencing of AURKA, TPX2, or their combination in SW480 and LOVO cells, whereas Bcl-2 expression was markedly decreased (P < 0.05) (Figure 6B). Furthermore, TPX2 knockdown or co-knockdown with AURKA markedly raised caspase 3 and p53 Levels compared to AURKA knockdown alone (P < 0.05) (Figure 6B). Similarly, in LOVO cells, TPX2 knockdown significantly increased Bax expression compared to AURKA knockdown; AURKA and TPX2 co-knockdown markedly increased Bax, caspase 3, and p53 expression (P < 0.01) (Figure 6B). The Transwell assay indicated a remarkable drop in cell migration and invasion abilities post AURKA or TPX2 knockdown (P < 0.01), with an even more pronounced reduction observed after dual knockdown (P < 0.05) (Figure 6C and D).

Figure 6 Knockdowns of Aurora A and targeting protein for Xklp2 promoted colorectal cancer cell apoptosis and inhibited cell migration and invasion. A: Cell apoptosis in SW480 and LOVO cells was determined by flow cytometry. Apoptosis cells were increased in Aurora kinase A (AURKA)/targeting protein for Xklp2 (TPX2) knockdown or co-knockdown cells; B: Western blotting was used to detect the expression levels of apoptosis-related biomarkers. B-cell lymphoma-2 (Bcl-2)-associated X protein, caspase 3, and tumor protein P53 expression levels were increased in SW480 and LOVO cells with AURKA/TPX2 knockdown or co-knockdown, while Bcl-2 was decreased; C: The number of migrated cells; D: Invaded cells was decreased in SW480 and LOVO cells with AURKA/TPX2 knockdown or co-knockdown. Transwell assay was used to detect cell migration and invasion abilities. ¹P < 0.05 vs negative control of short hairpin RNA (sh-NC) group. ²P < 0.01 vs sh-NC group. ³P < 0.05 vs short hairpin RNA of AURKA (sh-AURKA) group. ⁴P < 0.01 vs sh-AURKA group. ⁵P < 0.05 vs short hairpin-TPX2 (sh-TPX2) group. ⁶P < 0.01 vs sh-TPX2 group. AURKA: Aurora kinase A; TPX2: Targeting protein for Xklp2; sh-NC: Negative control of short hairpin RNA; sh-AURKA: Short hairpin RNA of Aurora kinase A; sh-TPX2: Short hairpin RNA of targeting protein for Xklp2; Bcl-2: B-cell lymphoma-2; Bax: B-cell lymphoma-2 associated X protein; p53: Tumor protein P53; FITC: Fluorescein isothiocyanate; PI: Propidium iodide.

Additionally, epithelial-mesenchymal transition (EMT) is a key pathway in cancer cell migration and invasion. N-cadherin and vimentin, EMT biomarkers, levels were significantly reduced in SW480 and LOVO cells after knockdown of AURKA, TPX2, or their combination (P < 0.01), whereas E-cadherin levels were increased (P < 0.01) (Figure 7A and B). Notably, co-knockdown of AURKA and TPX2 was more effective in inhibiting N-cadherin and vimentin expression and increasing E-cadherin expression (P < 0.05) (Figure 7A and B). In addition, a significant decrease in AURKA, phosphorylated AURKA, and TPX2 expression levels in SW480 and LOVO cells after knockdown of AURKA, TPX2, or both suggests the inhibition of AURKA/TPX2 activity (P < 0.01) (Figure 7A and B).

Figure 7 Knockdowns of Aurora A and targeting protein for Xklp2 inhibited epithelial-mesenchymal transition and knockdowns of Aurora A/targeting protein for Xklp2 pathways. A: Knockdowns of Aurora kinase A (AURKA)/targeting protein for Xklp2 (TPX2) or co-knockdown in SW480; B: AURKA/TPX2 knockdown or co-knockdown LOVO cells decreased N-cadherin, vimentin, AURKA, phosphorylated AURKA, and TPX2 expression, and increased E-cadherin expression. Epithelial-mesenchymal transition biomarkers include N-cadherin, E-cadherin, and vimentin. Western blotting was used to detect relative protein expression. ¹P < 0.05 vs negative control of short hairpin RNA (sh-NC) group. ²P < 0.01 vs sh-NC group. ³P < 0.05 vs short hairpin RNA of AURKA (sh-AURKA) group. ⁴P < 0.01 vs sh-AURKA group. ⁵P < 0.05 vs short hairpin-TPX2 (sh-TPX2) group. ⁶P < 0.01 vs sh-TPX2 group. AURKA: Aurora kinase A; TPX2: Targeting protein for Xklp2; sh-NC: Negative control of short hairpin RNA; sh-AURKA: Short hairpin RNA of Aurora kinase A; sh-TPX2: Short hairpin RNA of targeting protein for Xklp2; p-AURKA: Phosphorylated Aurora kinase A.

DISCUSSION

We analyzed three CRC-related gene databases, with AURKA and TPX2 emerging as the significant factors. They are highly expressed in the CRC tissues. Jung et al[9] found that high AURKA levels significantly enhanced the survival rates of patients with COAD. However, most studies reported that AURKA is correlated with CRC deterioration. Tang et al[18] determined that upregulation of AURKA promotes CRC development. AURKA associates with TPX2, a microtubule-associated factor, at the kinetochore, thereby recruiting it and stabilizing the microtubules to correct bipolar spindle formation. Asteriti et al[10] found that AURKA accumulation in interphase nuclei requested AURKA and TPX2 co-overexpression, suggesting that AURKA/TPX2 co-overexpression may contribute to tumorigenesis. Notably, TPX2 is crucial for the spatial regulation of spindle assembly, and its absence suppresses prostate cancer cell proliferation, attenuates tumorigenesis, and increases apoptotic rates[19]. Targeting TPX2 impairs CRC proliferation[20]. PPI network analysis showed that AURKA strongly interacts with TPX2. Most studies have shown that upregulation of AURKA or TPX2 inhibits CRC proliferation[21]. However, few studies have explored AURKA and TPX2 co-knockdown’s effects on the malignant phenotype of CRC. We found that AURKA and TPX2 co-knockdown enhanced the inhibition of silencing AURKA or TPX2 alone on CRC proliferation, invasion, and migration.

Interestingly, we found that the co-knockdown further enhanced cyclin B1 protein inhibition. Co-knockdown of AURKA and TPX2 may enhance the suppressive effect on CRC proliferation via cyclin B1. Major mitotic protein kinase CDK1-cyclin B1 phosphorylated importin-α1 to release TPX2 with promoting mitotic spindle assembly[22]. With reduced cyclin B1 activity, the cell cycle is arrested at the G2/M phase[23]. We found that knockdown of AURKA, TPX2, or their combination inhibited cyclin B1 expression during G2/M phase and decreased CDK2 expression. Furthermore, cyclin B1 overexpression enhances the progression of drug resistance in CRC[24]. The dual downregulation of AURKA and TPX2 could potentially improve the efficacy of CRC treatment.

We found that the inhibition of AURKA, TPX2, or their combination promoted apoptosis in CRC cells. Additionally, the co-knockdown of AURKA and TPX2 activated the Bax/Bcl-2 pathway. Our results supported the hypothesis that AURKA inhibitors promote apoptosis in CRC cells[25]. Bax and Bcl-2 are two important proteins in controlling apoptosis and are associated with the malignant phenotype of CRC. Bax promotes apoptosis, whereas Bcl-1 inhibits it. Studies have indicated that dysregulation of AURKA leads to DNA damage during mitosis, which is sensed in the subsequent G1 phase by a p53-dependent post-mitotic checkpoint[26]. Our results also support previous research showing that AURKA inhibition promotes the pro-apoptotic protein levels, such as p53 and Caspase-3[27]. Additionally, we found that inhibition of TPX2 or AURKA/TPX2 had a greater effect on apoptosis in CRC. This may be related to the reverse regulation of AURKA accumulation in interphase nuclei by TPX2 expression[10].

Importantly, cell migration and invasion abilities are crucial for CRC metastasis. To evaluate this ability, we observed that knocking down either AURKA, TPX2, or both noticeably reduced CRC cell invasion and migration ability. E-cadherin, N-cadherin, and vimentin are hallmarks of EMT and are correlated with poor CRC outcomes. AURKA promotes CRC cell migration and invasion via the EMT signaling pathway[28]. Although TPX2 promotes EMT in other malignancies[29], its effect on CRC EMT remains unclear. Here, we provide evidence for the regulation of EMT by TPX2. Interestingly, the combined decrease in AURKA and TPX2 expression enhanced the inhibitory effects of a single knockdown.

In summary, we demonstrated the inhibitory effects of AURKA and TPX2 on CRC malignant phenotype and that co-downregulation amplified these effects. However, this study had several limitations. The outcomes of cellular studies often deviate considerably from in vivo outcomes. Therefore, corroborating cellular evidence with in vivo experiments and extensive clinical data evaluation are essential to confirm the therapeutic role of co-depletion against tumorigenesis.

CONCLUSION

We analyzed the GEO gene database to identify potential key genes related to CRC. By constructing SW480 and LOVO cell models with AURKA and TPX2 knockdown via plasmid transfection, the inhibition of AURKA and TPX2 knockdown on CRC cell proliferation, migration, and invasion was observed. Furthermore, it has been demonstrated the co-downregulation enhanced these effects. This study provides the scientific basis for developing new treatment strategies and therapeutic drugs. Of course, there is still a gap between the results of this study and its clinical application, and more comprehensive in vivo and clinical observational experiments are needed to support this.