The gene expression profile of cholangiocarcinoma (GSE26566) was investigated through the Gene Expression Omnibus (GEO) database. Significant gene expression changes were identified, particularly those in molecular pathways involving DNA topoisomerase (ATP-hydrolyzing activity) activity in Gene Ontology (GO:2000371), through comparative and functional analyses. The databases were analyzed, and focus was placed on p-values of 0.01 and log2-transformed expression fold changes of $>$0.1.

The Chi Mei Medical Center enrolled 182 iCCA patients who underwent curative surgery between 1990 and 2010. Patients with metastatic disease or nodal metastases were excluded. This study was authorized by the Institutional Review Board (IRB) of The Chi Mei Medical Center with approval number 09912003. All participants provided their informed consent. Patients’ demographics and clinical details was collected retrospectively, including pathological characteristics, oncological survival follow-up, and cause of death. The Tumor, Node, Metastasis (TNM) system created by the eighth edition American Joint Committee on Cancer (AJCC) in 2017 was used to measure the tumor stage. Two pathologists examined the tumor samples.

The formalin-fixed tissues were embedded in paraffin and sectioned into 10 $\mu$m sections. Subsequently, immunohistochemical (IHC) staining of TOP2A was conducted using a primary monoclonal antibody directed towards TOP2A (1:200, EP1102Y, Epitomics, Cambridge, UK) for a duration of 1 hour. To detect the primary antibodies, the ChemMate EnVision kit (DAKO, K5001, Carpinteria, CA, USA) was utilized. The secondary antibody was incubated with the slides for 30 minutes and then developed with 3,3-diaminobenzidine for 5 minutes. Subsequently, the slides were counterstained with hematoxylin. Cholangiocarcinoma expressing TOP2A was utilized as a positive control, while rabbit serum IgG replaced the primary antibody as a negative control. Two pathologists used the formula H-score = SPi (i + 1) to compute the H-score, where Pi stands for the percentage of stained tumor cells in different intensities ranging from 0% to 100% and i for the degree of staining (0 to 3+). The two pathologists evaluated the slides concurrently and decided on an H-score in case there were any scoring disagreements. Based on the median H-score, the immunostaining was categorized as having low or high expression levels.

The relative expression and knockdown efficacy of the TOP2A gene in iCCA cell lines were determined using RT-PCR, as previously published [19]. Total RNA was collected from iCCA cell lines, and real-time RT-PCR was employed to determine the transcription level of TOP2A. The mRNA abundance of TOP2A (Hs01032137_mL) was assessed using predesigned TaqMan assay reagents (Cat. No. Hs01032137_mL, Applied Biosystems, Waltham, MA, USA) and the ABI StepOnePlusTM system (Applied Biosystems, Waltham, MA, USA), with POLR2A (Hs01108291_mL) used as the internal control for normalization.

The SNU1079 and SNU1196 cell lines were procured from a cell bank based in Seoul, South Korea. Initially, these cell lines were cultured in ACL-4 medium supplemented with 5% heat-inactivated fetal bovine serum. After establishment, the cell cultures were maintained in RPMI 1640 medium and supplemented with 10% heat-inactivated fetal bovine serum. These cells were cultured under controlled conditions in a humidified incubator at 37 °C in the presence of 5% CO${}_2$ and 95% air as previously described [20]. On the other hand, the Huh28 and HuCCT1 cell lines were obtained from a cell bank in Osaka, Japan. These cell lines were cultured in either Roswell Park Memorial Institute 1640 medium (HuCCT1) or minimal essential medium (Huh28), which were both supplemented with 100 U/mL penicillin, 100 $\mu$g/mL streptomycin, and 10% fetal bovine serum as previously reported [21]. All the cell lines were authenticated by short tandem repeat genotyping (ThermoFisher, Waltham, MA, USA), periodically confirmed to be mycoplasma free using PlasmoTest (Invivogen, San Diego, CA, USA).

The SNU1196 and Huh28 cell lines were modified to create stable TOP2A-silenced clones using lentiviral vectors pLKO.1-shLacZ (TRCN0000072223: 5′-TGTTCGCATTATCCGAACCAT-3′) and pLKO.1-shTOP2A (#1: TRCN0000049280: 5′-GCTCCAAATCAA TATGTGATT-3′; #2: TRCN0000049278: 5′-GCCCAA GTGTTCTTTAGCTTT-3′), which were obtained from the Taiwan National RNAi Core Facility in Taipei, Taiwan. These cell lines initially had high TOP2A expression that was reduced by shRNAs against TOP2A (shTOP2A). To generate viruses for the modification, HEK293 cells were transfected with the three vectors mentioned above, using Lipofectamine 2000 from Thermo Fisher Scientific in Waltham, MA, USA, as previously described [19].

Primary antibodies against TOP2A (clone AA6, 1:500; Millipore, Beverly, MA, USA) were utilized, and a previously reported western blotting technique was utilized to assess TOP2A expression and the effect of TOP2Aknockdown in SNU1196 and Huh28 cell lines. Glyceraldehyde 3-phosphate dehydrogenase (GADPH) was used as a control for protein loading (6C5, 1:10,000; Millipore, Beverly, MA, USA) [22]. To immobilize the protein, cell lysates containing 25 $\mu$g of protein were separated on a 4% to 12% gradient NuPAGE gel (Invitrogen, Carlsbad, CA, USA) and transferred to polyvinylidene difluoride membranes (Amersham Biosciences, Buckinghamshire, UK). To prevent non-specific binding, membranes were blocked with TBST buffer and 5% skimmed milk at room temperature for one hour. Primary antibodies were then exposed to the membranes at 4 °C overnight. After that, the membranes were incubated with the secondary antibody for 1.5 hours at room temperature, and proteins were identified using a chemiluminescence system (Amersham Biosciences, Buckinghamshire, UK).

The enzyme-linked immunosorbent assay (ELISA)-based and colorimetric bromodeoxyuridine (BrdU) assay (Roche Holding AG, Basel, Switzerland) was utilized to quantify DNA synthesis. At 24, 48, and 72 hours, the amount of DNA synthesis was measured in the TOP2A-knockdown or shLacA control SNU1196 and Huh28 cell lines. After three hours of BrdU incubation at 37 °C and 5% CO${}_2$, the labeling medium was removed and the cells were fixed before being incubated with an anti-BrdU-POD solution. Using an ELISA reader (Promega Corp., Madison, WI, USA), the absorbance at 450 nm was measured, with the reference absorbance set to 690 nm.

The experimental protocol for cell migration and invasion was carried out in accordance with a previously documented procedure [22]. For the cell invasion experiment, the 24-well Collagen-based Cell Invasion Assay from Millipore, Beverly, MA, USA, and Falcon HTS FluoroBlok 24-well inserts from BD Biosciences, Franklin Lakes, NJ, USA, were used. The inserts were rehydrated using serum-free medium and then placed in the upper chamber, which contained a serum-free suspension with an equal number of cells. Over a 12- to 24-hour incubation period, the cells were allowed to migrate towards the lower chamber, which contained a medium with 10% fetal bovine serum. Following removal of the non-invading cells from the upper chamber, the invading cells were stained, lysed in extraction buffer, and then transferred to 96-well plates for 560 nm colorimetric readings.

In order to determine the unknown functions of TOP2A in iCCA, the transcription level of TOP2A and its coexpressed genes contained in the cholangiocarcinoma dataset (n = 51, Firehose Legacy, TCGA) were analyzed to establish correlations. Subsequently, the top 200 differentially expressed genes with positive or negative correlations with TOP2A were selected for functional annotation using the Gene Ontology (GO) classification system and rated by fold enrichment.

The statistical analyses were conducted using SPSS software (version 28, IBM Corp., Chicago, IL, USA). The Chi-square test was utilized to evaluate the association between TOP2A expression stage and clinicopathologic characteristics. Local recurrence-free survival (LRFS), metastasis-free survival (MFS), and disease-specific survival (DSS) were calculated from the beginning of therapy to the date of the event, and the latest follow-up date was noted for patients who were lost to follow-up during the study period. Kaplan-Meier analysis was employed to generate survival curves, and log-rank tests were utilized to identify prognostic differences between groups. All analyses were performed using two-sided significance tests, and a p-value less than 0.05 was considered significant.

We mined the public CCA transcriptome dataset (GSE26566) and compared the results with DNA topoisomerase (ATP-hydrolyzing activity) activity (GO:2000371). We discovered that five genes were significantly associated with DNA topoisomerase (ATP-hydrolyzing activity) activity (Table 1), and TOP2Awas found to be a high-ranking differently expressed candidate gene that showed a significant difference (p 0.0001) and a log-ratio of $>$0.1 (Table 1). Moreover, the heat map data mining results also showed that expression of DNA topoisomerase-related genes differed between CCA tissue and the surrounding liver and non-tumor biliary epithelium tissues, and TOP2A was shown to be highly expressed in CCA patients (Fig. 1). The above data demonstrate that TOP2A might play a critical role in cancer progression in CCA patients.

Fig. 1.

Data mining of genes expression compared between cholangiocarcinoma, surrounding liver and non-tumor biliary epithelium from the public domain (GSE26566).

Our previous results indicate that high expression of TOP2A might be correlated with CCA progression. We further analyzed associations between TOP2A expression and the clinicopathologic characteristics of CCA patients. We included data from 182 iCCA patients, 108 males and 74 females, where 107 patients were older than 65 and 75 patients were younger than 65. These cases were classified as 105 of the large duct type and 77 of the small duct type based on histological variants; 61 cases were classified as well-differentiated, 66 as moderately differentiated, and 55 as poorly differentiated according to histological grades. A total of 163 cases were classified as R0 and 19 cases were classified as R1 based on the surgical margin, and in terms of the primary tumor size, 87 cases were classified as T1, 61 as T2, and 34 as T3 (Table 2). Taken together, our results indicate that TOP2A overexpression is highly correlated with the co-occurrence of hepatitis (p-value = 0.002), the primary tumor size (p-value = 0.001), and the histological variants (p-value = 0.002) of clinicopathological parameters in cholangiocarcinoma patients with iCCA (Table 2). Moreover, we determined that TOP2A is overexpressed in cholangiocarcinoma patients (Fig. 1). We confirmed this by IHC staining of cholangiocarcinoma patient tissue sections (Fig. 2). Moreover, we calculated the immunohistochemical staining H-score which showed significantly higher in iCCA cells than hepatocytes and cholangiocytes (Fig. 3). These data suggest that TOP2A expression is significantly associated with clinicopathological variables in iCCA patients.

Fig. 2.

Identify the TOP2A expression in iCCA tissues. (A) In low-stage tumor tissues, TOP2A positive cells are weak detected (a–d). (B) In high-stage tumor tissues, tumor cells showed positive TOP2A staining (e–h). (C) A comparison of H-score showed significantly higher TOP2A expression in iCCA cells than hepatocytes and cholangiocytes (N = 182 in each group).

Fig. 3.

Kaplan–Meier survival curves in the CCA patients according to surgical margin, primary tumor staging and TOP2A expression. R0 Resection, pT1 stage and low TOP2A expression correlated with better overall survival (A,B,C), better disease-specific survival (D,E,F), better local recurrence-free survival (G,H,I) and better metastasis-free survival (J,K,L).

To determine whether TOP2A expression is correlated with CCA patients’ survival outcomes, we analyzed the association between TOP2A expression and the overall survival rate, disease-specific survival rate, local recurrence-free survival rate, and metastasis-free survival rate in CCA patients. For overall survival, an R0 stage surgical margin, T1 stage primary tumor, and low TOP2A expression (Fig. 3A–C) were associated with better outcomes in both the univariate and multivariate analyses (Table 3). The dis-ease-specific survival endpoints analysis showed the same results whereby R0 stage, T1 stage, and low TOP2A expression patients had better outcomes (Table 3) (Fig. 3D–F) than other patients. Moreover, patients with a T1 stage primary tumor, large duct type histological variant, and histological low grade were shown to have better outcomes in the univariate analysis of local recurrence-free survival (Table 4). Importantly, an R0 stage surgical margin, T1 stage primary tumor, and low TOP2A expression were all significantly associated with better outcomes (Fig. 3G–I) in both the univariate analysis and the multivariate analysis of local recurrence-free survival and metastasis-free survival (Table 4) (Fig. 3J–L).

Taken together, our results indicate that TOP2A is highly expressed in CCA patients, and high TOP2A expression is correlated with poor prognosis and notably worse overall survival (p = 0.0011, Fig. 3C), disease-specific survival (p 0.0001, Fig. 3F), local recurrence-free survival (p 0.0001, Fig. 3I), and metastasis-free survival (p 0.0001, Fig. 3L) rates.

To determine the unknown functions of TOP2A in IHCC, a set of the top 200 differentially expressed genes with positive (Supplementary Table 1) or negative correlations (Supplementary Table 2) with TOP2A were downloaded from the cholangiocarcinoma dataset (n = 51, Firehose Legacy, TCGA). These genes were then utilized for a functional annotation analysis using the Gene Ontology (GO) classification system. The biological processes most positively correlated with TOP2A were positive regulation of DNA endoreduplication (GO: 0032877, fold enrichment: $>$100), cell cycle DNA replication maintenance of fidelity (GO: 1902298, fold enrichment: $>$100), and mitotic DNA replication maintenance of fidelity (GO: 1990505, fold enrichment: $>$100) (Fig. 4A). DNA repair protein RAD51 homolog 1 (RAD51) and breast cancer type 2 susceptibility protein (BRCA2) genes, involved in most of the biological processes mentioned above, were also identified. As for molecular functions, the most positively correlated term with TOP2A was D-loop DNA binding (GO: 0062037, fold enrichment: 60.85) (Fig. 4B), and the RAD51-associated protein 1 (RAD51AP1) gene was identified.

Fig. 4.

The characteristic GO terms enriched in TOP2A upregulation. The top 200 differentially expressed genes with positive relationship to TOP2A were annotated utilizing the GO classification system depending on (A) biological processes and (B) molecular functions and were rated by fold enrichment.

In order to assess the impact of TOP2A, we initially measured the levels of endogenous TOP2A expression in four different cholangiocarcinoma cell lines. We found that SNU1196 and Huh28 cells had the highest amounts of TOP2A transcripts and protein expression (Fig. 5A). We then used short hairpin RNA (shRNA) to successfully knock down TOP2A in both SNU1196 (Fig. 5B, left) and Huh28 (Fig. 5B, right) cell lines. TOP2A-silenced SNU1196 (Fig. 5E, left) and Huh28 (Fig. 5E, right) cells had considerably reduced proliferation (viability). We also looked at TOP2A’s role in cholangiocarcinoma cell migration and invasion. TOP2A knockdown significantly reduced SNU1196 (Fig. 5C, left and Fig. 5D, left) and Huh28 (Fig. 5C, right and Fig. 5D, right) cell migratory and invasive abilities.

Fig. 5.

TOP2A expression promotes the growth of CCA cells in vitro. (A) SNU1196 and Huh28 cells had the highest amounts of TOP2A mRNA and protein expression among four CCA cell lines. (B) The two cell lines with high endogenous TOP2A expression were stably silenced by a lentiviral vector bearing one of the two clones of TOP2A shRNA with different sequences for both SNU1196 (left panel) and Huh28 (right panel) cells. (C) Cell migration ability was significantly reduced in TOP2A- knockdown SNU1196 and Huh28 cell lines. (D) Cell invasion ability was significantly reduced in TOP2A- knockdown SNU1196 and Huh28 cell lines. (E) Using an ELISA-based colorimetric assay to assess the rate of BrdU uptake, cell proliferation was significantly reduced in TOP2A-knockdown SNU1196 and Huh28 cell lines (* p 0.05, ** p 0.01).