Fresh tissue (n = 6) with histologically confirmed cancer was obtained from LUAD patients who had not undergone any prior treatments and who were admitted to the Fourth Hospital of Hebei Medical University.

The Oncomine database is a powerful set of analysis functions that compute gene expression signatures, clusters and gene-set modules, thus automatically extracting biological insights from 715 datasets and 86,733 samples (https://www.oncomine.org/resource/login.html) [19]. The expression levels of TACC3 in LUAD were determined here by Oncomine Database analysis. The following threshold was considered as statistically significant: p = 0.0001, fold change = 1.5, gene rank = all.

GEPIA (http://gepia.cancer-pku.cn/) is a newly developed interactive web server for analyzing RNA sequence expression data for 9736 tumor and 8587 normal samples from the TCGA and GTEx projects using a standard processing pipeline [20]. TACC3 expression in LUAD was determined in the single gene expression analysis using the following thresholds: $|$log${}_2$$|$FC cutoff = 0.5 and p-value cutoff = 0.01. In the survival analysis, the threshold was determined according to the following values: group cutoff: median; cutoff-high (%) and cutoff-low (%): 50 and 50.

UALCAN is a comprehensive and interactive web resource (http://ualcan.path.uab.edu/index.html) designed to provide easy access to publicly available cancer OMICS data (TCGA and MET500), thus facilitating the identification of candidate biomarkers. The expression of TACC3 and patient survival data were selected from the TCGA databases, with p 0.05 considered significant.

Gene expression and patient data for LUAD were downloaded from Genomic Data commons (GDC) (https://portal.gdc.cancer.gov/) using the GDC data transfer tool. Gene expression data were analyzed by R (version: 3.6.1, Boston, MA, USA) with related R packages. Clinical parameters such as age, gender, survival status, and tumor stage were extracted from the patient data and matched to each patient using the PERL script.

Based on a meta-analysis, the Kaplan Meier plotter (http://kmplot.com/analysis/) is capable of assessing the association of 54,000 genes with patient survival in 21 cancer types [21, 22]. The correlation between TACC3 expression and overall survival (OS) in LUAD, as well as with First Progression (FP), was analyzed using mRNA data. Patients were divided according to the median value, with the cutoff value set as auto select.

To assess the diagnostic value of TACC3, ROC curves were generated using SPSS 26 (IBM SPSS Statistics, Chicago, IL, USA). The AUC was determined and is shown for each panel.

To identify potential mechanisms that underlie the associations of TACC3 expression in LUAD, KEGG analysis was performed to determine whether an a priori defined set of genes was differentially expressed between high and low TACC3 expression groups [23]. Gene sets with a normal p-value of 0.05 and false discovery rate (FDR) of 0.05 were considered to be significantly enriched.

qRT-PCR was performed to determine the expression of TACC3 mRNA. Briefly, total RNA from the surgically obtained paired (tumor and normal) tissues (n = 6) was isolated using the TRI Reagent RNA Isolation Reagent (Sigma-Aldrich) according to the manufacturer’s instructions. The first-strand template complementary DNA (cDNA) was synthesized using a reverse transcription kit (Takara, Japan). The primer sequences were as follows: TACC3: 5’-CCTCTTCAAGCGTTTTGAGAAAC-3’ (sense) and 5’-GCCCTCCTGGGTGATCCTT-3’ (antisense); $\beta$-actin: 5’- CGCGAG AAGATGACCCAGAT-3’ (sense) and 5’-GGGCATACCCCT CGTAGATG-3’ (antisense) [10, 18]. $\beta$-actin was used as the internal control. The PCR reaction was performed as follows: denaturation at 94 °C for 3 minutes, and then 35 cycles of 94 °C for 30 seconds, 56 °C for 30 seconds, 72 °C for 2 minutes, and finally an elongation step at 72 °C for 10 minutes. Three independent experiments were performed.

All analyses of the TCGA data were conducted using R software (version 3.6.1). Univariate Cox regression analysis was used to select potential prognostic factors, and multivariate Cox analysis was performed to determine the correlation between TACC3 expression and survival, as well as with other clinical features. Statistical analyses were performed using SPSS 26 (IBM SPSS Statistics, USA), with p 0.05 considered to be statistically significant.

Clinical data for 467 patients was downloaded from the TCGA database and included age, gender, stage, TMN classification and survival status (Table 1).

TACC3 mRNA expression in LUAD was determined from the Oncomine, GEPIA and UALCAN databases. The level of TACC3 expression was higher in LUAD tissues compared with normal tissues (Fig. 1A–C). These results were further confirmed by analysis of the TCGA database (p = 2.672e-11, Fig. 1D). In addition, a significant difference in TACC3 expression was observed in a paired comparison of LUAD with adjacent normal tissues (p = 2.343e-15, Fig. 1E). These results demonstrate that TACC3 was highly expressed in LUAD compared with normal tissues.

Fig. 1.

TACC3 expression levels in LUAD cancer determined using four databases. (A) Oncomine database. Cell color is determined by the best gene rank percentile for the analyses within the cell. (B) GEPIA database. Significant differences between cancer types are shown in red, p 0.05. (C) UALCAN database. Normal vs tumor, p = 1.6235e-12. (D) TCGA database analyzed by R script. p = 2.672e-11. (E) TACC3 expression in a paired comparison of LUAD and adjacent normal tissues. Data was extracted from the TCGA database, p = 2.343e-15.

The prognostic value of TACC3 expression in LUAD patients was determined using the Kaplan-Meier Plotter, GEPIA, UALCAN, and TACG databases. No significant difference was observed for disease free survival (DFS) with GEPIA (Log rank p = 0.05, HR = 1.4; Fig. 2D). However, high TACC3 mRNA expression was significantly associated with poor OS of LUAD patients in Kaplan-Meier Plotter (Log rank p = 1.7e-07, HR = 1.88; Fig. 2B), UALCAN (p 0.0001; Fig. 2C), GEPIA (Log rank p = 1e-04, HR = 1.8; Fig. 2E), and TCGA (p = 0.002; Fig. 2F) when analyzed by R. In addition, high TACC3 mRNA expression in LUAD patients was significantly associated with worse FP in Kaplan-Meier Plotter (Log rank p = 5.3e-05, HR = 1.91; Fig. 2A). These results indicate that TACC3 expression has a major impact on the survival of LUAD patients.

Fig. 2.

Kaplan-Meier survival curves for TACC3 expression in LUAD as determined using four databases. (A,B) OS and RFS from Kaplan-Meier Plotter database. (C,D) OS and DFS from GEPIA database. (E) OS from UALCAN database. (F) OS from TCGA database. OS, Overall Survival; RFS, Relapse-Free Survival; DFS, Disease Free Survival.

The relationship between TACC3 expression and the survival of various clinicopathological subgroups of LUAD patients was determined using Kaplan-Meier plotter. As shown in Table 2, TACC3 mRNA expression was negatively associated with OS in both female and male patients (p = 0.0084, p = 9.1e-4, respectively), as well as in LUAD patients who were smokers (p = 0.008). High expression of TACC3 was also associated with worse OS in stage 1, 3, AJCC stage T1, stage N0, and M0 patients (p = 9.7e-5, p = 0.023, p = 5.2e-4, p = 0.0035 and p = 7.6e-4, respectively). TACC3 expression was also associated with FP in males, non-smokers, and stage 1 LUAD patients (p = 5.1e-4, p = 0.027 and p = 0.0058, respectively). No significant associations were observed between TACC3 expression and the PPS of LUAD patients. These results suggest that TACC3 expression is mainly associated with OS in LUAD patients.

ROC analysis was performed to assess the diagnostic value of TACC3 expression in LUAD patients. The observed AUC of 0.940 suggests that TACC3 mRNA expression has strong diagnostic value for LUAD (Fig. 3A). We further examined the diagnostic value of TACC3 in stage I, II, III, and IV LUAD patients (Fig. 3B–E). These showed AUCs of 0.930, 0.956, 0.973, and 0.904, respectively, indicating that TACC3 is a potential diagnostic biomarker for LUAD.

Fig. 3.

Diagnostic value of TACC3 expression in LUAD using ROC analysis. (A) ROC curve for TACC3 mRNA expression in normal and LUAD tissues. (B–E) ROC curves for TACC3 mRNA expression in different stages (I, II, III, IV) of LUAD. All AUCs are shown for each ROC curve. AUC, area under ROC curve.

We next performed univariate and multivariate Cox regression analyses using R script to determine whether TACC3 expression was an independent risk factor for OS in LUAD patients. Stage, T and N classification, and TACC3 expression were all independent risk factors for OS in univariate Cox analysis (p = 3.07e-11, 2.24e-5, 7.17e-9, and 8.23e-3, respectively; Table 3). In multivariate Cox analysis, only stage and TACC3 expression were identified as independent risk factors (p = 2.2e-4 and p = 4.1e-4, respectively). These results indicate that TACC3 expression is an independent risk factor for worse OS in LUAD patients (HR = 1.43, 95% CI: 1.17– 1.7, p 0.001; Fig. 4).

Fig. 4.

Multivariate Cox regression analysis shows that TACC3 is an independent risk factor for OS among LUAD patients. **p 0.01, ***p 0.001.

Real-Time PCR was performed to determine TACC3 mRNA levels in 6 pairs of matched LUAD tissues and their adjacent non-cancerous tissues. The expression of TACC3 mRNA was upregulated in LUAD cancer tissues compared to adjacent noncancerous tissues (p 0.01, Figs. 5,6).

Fig. 5.

Enrichment plots for LUAD cases with high TACC3 expression using GSEA 3.0. GSEA results showing differential enrichment of genes related to various pathways. (A) Cell cycle. (B) Spliceosome. (C) Homologous recombination. (D) Progesterone-mediated oocyte maturation. (E) Oocyte meiosis. (F) Ubiquitin-mediated proteolysis. (G) p53 signaling pathway. (H) Notch signaling pathway. (I) Proteasome.

Fig. 6.

TACC3 mRNA expression is upregulated in LUAD tissue compared to adjacent normal tissue. T: LUAD tissues; N: noncancerous tissues. The results shown are the mean $\pm$ SD from six paired samples. **p 0.01.