Background

Renal cell carcinoma (RCC) is the most common kidney cancer that occurs in adulthood [1, 2]. RCC is classified into different subtypes based on morphology and histology, including clear cell, papillary cell carcinoma, chromophobe, oncocytoma, collecting duct, and unclassified. Clear cell RCC is the most common subtype, accounting for 75% of RCCs. As the cancer itself is rather resistant to chemotherapy or radiotherapy, surgical removal is the only curative treatment available to date [1, 3]. However, approximately 40% of patients undergoing treatment for primary tumors will relapse with local or metastatic RCC with a 5-year survival rate of 60–70% in organ-confined tumors versus 10% for metastatic tumors [3,4,5,6]. Reports indicated that the incidence of subsequence malignancies varies by population level, ranging from 4.5 to 26.9%, with the most common subsequent malignancies following an RCC diagnosis being prostate, bladder, lung, breast, and colon cancers [7,8,9]. Analyses from The Cancer Genome Atlas (TCGA) indicated that RCC histological subtypes were associated with distinct genetic and epigenetic changes that contributed to the pathogenesis of RCC. The most common genetic alteration associated with clear cell RCC is the Von Hippel‒Lindau (VHL) gene, which is known to alter the half-life of hypoxia-induced factor (HIF) alpha and consequently promotes a prominent angiogenesis process that supplies blood to the tumor [10, 11]. Additionally, VHL hypermethylation is characteristic of most clear cell RCC [12]. It has been proposed that increased DNA hypermethylation was associated with disease at a higher stage and lower survival in all RCC histopathological subtypes [12]. The discovery of particular molecular biomarker associated with survival will enhance patient treatment strategies and offer potential targets for precision therapy designed for specific subtypes. Over the last two decades, medical treatment for RCC has transitioned from the use of cytokines to targeted therapies against vascular endothelial growth factor (VEGF) or tyrosine kinase to combination therapies involving immune checkpoint inhibitors and tyrosine kinase inhibitors [3, 13]. However, there are still many unknown mechanisms that might be involved in RCC pathogenesis that should be further investigated.

The transcription factor (TF) activator protein (AP)-2α, also known as TFAP2A, represents the best-characterized member of the AP-2 TF family that plays important roles in several biological processes, such as cell growth, apoptosis, and tissue differentiation during early development [14, 15]. Oya et al. showed that distinct expression patterns of AP-2 isoforms in RCC, with AP-2α predominantly expressed in clear cell RCC [16]. Aberrant AP-2α expression has been observed in various cancers, acting as a tumor suppressor or oncogene depending on the cancer type [17,18,19,20,21,22,23]. Despite its recognized involvement in different cancers, the role of AP-2α in RCC remains unexplored. While epigenetic alterations of the AP-2α gene, where it function as a tumor suppressor, have been reported in breast cancer and melanoma [17, 24], no study to date has demonstrated similar DNA methylation regulatory mechanisms in the context of RCC. Moreover, there has been a lack of comprehensive exploration into the biological role, prognostic value, and clinical significance of the AP-2α gene in RCC. Consequently, this study aimed to investigate the clinical significance of the expression of AP-2α in RCC and determine whether epigenetic regulation of the AP-2α gene plays an important role in the carcinogenesis of RCC.

Methods

Patients and specimens

A total of 107 RCC samples were collected from the tissue bank between 2004 and 2010, and this study was approved by the Institutional Review Board of Chang Gung Memorial Hospital, LinKou (IRB no. 100-4590B). Among them, 27 RCC samples were used for the Methylated-CpG island recovery assay (MIRA) combined with microarray analysis to identify methylated genes. 80 RCC samples were used for the other experiments, including combined bisulfite restriction analysis (COBRA) and reverse transcription-quantitative polymerase chain reaction (RT‒qPCR). After surgical resection, tissue specimens were washed with PBS, snap-frozen in liquid nitrogen and stored at -80 °C. Tumors were staged according to the American Joint Committee on Cancer (AJCC) TNM guidelines, and the nuclear grade of the tumors was determined according to the Fuhrman grading system [25, 26]. All clinicopathological features are listed in Table 1. Follow-up data were obtained from patient health records and the cancer registration unit at Chang Gung Memorial Hospital.

Table 1 Association between AP-2α mRNA expression and clinicopathological features in RCC patients
Full size table

DNA and RNA preparation

DNA from tissues was extracted using a standard proteinase K/phenol/chloroform procedure. DNA was quantified using an ND-1000 spectrophotometer (NanoDrop Technologies, Inc., Wilmington, DE, USA) and then further analyzed by 1% agarose gel electrophoresis (Invitrogen Life Technologies, Carlsbad, CA).

Total RNA from tissues was extracted using TRIzol reagent (Invitrogen Life Technologies) and stored at -80 °C according to the manufacturer’s instructions. Total RNA was quantified and its quality was assessed using ND-1000 spectrophotometry.

Methylated-CpG island recovery assay and microarray analysis

To detect the DNA methylation pattern among different types of RCC, MIRA combined with microarray analysis was employed following the protocol described by Rauch et al. [27, 28]. Nimblegen’s Signalmap program (version 1.9, Roche-NimbleGen, Madison, WI) was used to visualize DNA methylation data and generate profiling snapshots.

Combined bisulfite restriction analysis

A total of 67 RCCs from 80 patients were included in the COBRA experiment. Of these, 13 cases had limited amounts of tissue and could only be used for RNA extraction. The bisulfite conversion of 500 ng DNA was performed using the EZ DNA methylation Gold kit (Zymo Research, Orange, CA, USA) according to the manufacturer’s instructions. The methyl primers obtained from Methyl Primer Express® Software v1.0 (Applied Biosystems Inc., Foster City, CA, USA) in COBRA were AP-2α methyl-1 forward primer (5’-GGTATTTTTTTTGGGGTAGGTA-3’) and reverse primer (5’-TACAACCTAAACCCCCTACAC-3’). The PCR products were digested with BstUI (New England Biolabs, Ipswich, MA, USA) at 60 °C for 90 min and separated on a 10% nondenaturing polyacrylamide gel. CpGenome universal methylated DNA (Merck Millipore, Bedford, MA, USA) was used as a positive control.

Reverse transcription-quantitative polymerase chain reaction

Complementary DNA (cDNA) synthesis was performed using a SuperScript™ III First-Strand Synthesis SuperMix kit (Invitrogen Life Technologies) according to the manufacturer’s instructions, and the product was stored at -20°C. qPCRs were performed using iQ™ SYBR® Green Supermix (Bio-Rad, Hercules, CA, USA) in a total volume of 20 µl with a Bio-Rad iQ5 iCycler Real-Time PCR Detection System following the manufacturer’s instructions. The primers used for amplification of AP-2α were as follows: forward primer (5’-GATCCCAATGAGCAAGTGAC-3’) and reverse primer (5’-ATGAGGTTGAAGTGGGTCAA-3’). In addition, a β-actin forward primer (5’-TCAGCAAGCAGGAGTATG-3’) and reverse primer (5’-GTCAAGAAAGGGTGTAACG-3’) were used. The -ΔCT value of relative AP-2α mRNA expression was calculated according to the formula: -ΔCT = -(CTTarget –CTInternal). CTTarget is the CT value of the target gene AP-2α in RCC tumor tissue samples, and CTInternal is the CT value of the internal control gene β-actin in RCC tumor tissue samples.

Bioinformatic analysis of the NCBI database

Data on AP-2α expression in TCGA/GTEx dataset and its association with overall survival and disease-free survival were derived from the GEPIA2 database (http://gepia2.cancer-pku.cn) [29].

Statistics

Statistical analysis was accomplished using the SPSS statistical software package (version 26.0; SPSS). One-way ANOVA was employed to compare AP-2α mRNA expression between different subtypes of RCC. Spearman rank correlation analysis was performed to examine the correlation between AP-2α mRNA expression levels and methylation status. The correlation coefficient (r) was calculated to measure the correlation degree between AP-2α mRNA expression level and methylation status. The association between AP-2α mRNA expression and clinicopathologic variables was estimated using Pearson’s chi-square test or Fisher’s exact test. We used the median -ΔCT value of -12.8 and − 13.1 as the cutoff value of AP-2α mRNA expression in patients with RCC and clear cell RCC respectively. This value was used to divide the samples into two groups: low expression (-ΔCT ≤-12.8 or -ΔCT ≤-13.1 in patients with RCC and clear cell RCC respectively) and high expression (-ΔCT >-12.8 or ΔCT >-13.1 in patients with RCC and clear cell RCC respectively). Survival curves were calibrated from the date of surgery until death related to RCC or the last follow-up finding. Disease-free survival curves were calibrated from the date of surgery until recurrence or metastasis associated with RCC or the last follow-up finding. Kaplan‒Meier curves were calibrated for each relevant clinicopathological variable and AP-2α mRNA expression; differences in survival time between subgroups of patients were analyzed by the log-rank test. Cox regression analysis was conducted at both univariate and multivariate levels. Differences were considered to be statistically significant at p < 0.05.

Results

Methylation status of AP-2α in RCC

From our methylated CpG island recovery assay in various RCC histological types (19 clear cell RCCs, 4 papillary RCCs, and 4 chromophobe RCCs), we identified 2806 genes with altered methylation, with 120 genes, including AP-2α, overlapping among the types (Fig. 1A). NimbleScan (version 2.3.78, NimbleGen Systems Inc.) analysis highlighted a frequently methylated region near exon 2 of the AP-2α gene (Fig. 1B). We verified this with the COBRA assay on 67 RCC clinical samples, revealing 22 CpG dinucleotides and 3 BstUI cutting sites in the methyl-primer amplified region (Fig. 1C). BstUI digested products appeared in the representative methylated #11 tumor samples but not in the unmethylated #8 tumor samples (Fig. 1D). The AP-2α gene was methylated in most clear cell (78.7%, 37/47) and papillary (50.0%, 4/8) RCC clinical tumor samples, but not in chromophobe (0%, 0/9) or other RCC (0%, 0/3) samples (Fig. 1E).

Fig. 1
figure 1

COBRA methylation analysis of the AP-2α gene in different subtypes of RCC. A, Venn diagram depicted the number of genes identified as methylated for 19 clear cell RCC, 4 papillary RCC and 4 chromophobe RCC. B, Snapshot views from Signalmap of AP-2α gene showed the transcription start site (TSS, position:10,523,256), CpG islands (gray column) and methylation area (black column) of these regions from the RCC samples. The data showed the methylation status of the representative 2 clear cell RCC, 2 chromophobe RCC, 2 papillary RCC, and 2 oncocytoma RCC. The arrow indicated the transcription direction of AP-2α gene. C, Schematic representation of CpG distribution in the exon 2 region of the human AP-2α gene. A total of 22 CpG dinucleotides were analyzed in the indicated region (vertical tick mark). Bisulfite PCR primers (arrows) were designed with Methyl Primer Express® Software v1.0 and used to amplify DNA from different subtypes of RCC tumor specimens. 3 BstUI cutting sites were included in the amplified fragment. D, Methylation analysis by COBRA for different paired RCC tumor (T) and adjacent nontumor tissue samples (N) was performed, and the representative results are shown. We used CpGenome universal methylated DNA as a positive control (C) for methylation studies. Arrows indicate unmethylated (U)/methylated (M) alleles. E, The methylation status of different subtypes of RCC was analyzed using the COBRA approach. 47 clear cell, 8 papillary, 9 chromophobes, and 3 other (2 granular and 1 unclassified) RCC clinical tumor samples were included in this experiment

Full size image

AP-2α gene expression in RCC

RT‒qPCR assays were performed to examine AP-2α mRNA expression in 80 RCC clinical samples. AP-2α mRNA expression was significantly lower in clear cell and papillary RCC compared with chromophobe and other RCC (p < 0.0001, Fig. 2A). To further investigate whether AP-2α mRNA expression was associated with methylation status as determined by the COBRA assay, Spearman rank correlation analysis was performed. The results revealed a trend of a negative correlation between AP-2α mRNA expression levels and methylation status (r = -0.384, p = 0.001, Fig. 2B).

Fig. 2
figure 2

Expression of AP-2α in different RCC subtypes. A, Relative AP-2α expression levels in different subtypes of RCC were determined by RT‒qPCR. The -ΔCT value of relative AP-2α expression was calculated according to the following formula: -ΔCT = -(CTTarget – CTInternal). B, The correlation between AP-2α mRNA expression levels and methylation status (p < 0.05 was considered to indicate significant)

Full size image

Association of AP-2α expression with clinicopathologic characteristics in RCC

Next, we investigated the association between AP-2α mRNA expression and clinicopathologic characteristics in RCC patients. The median -ΔCT value of -12.8 was used as the cutoff value. No significant association between the AP-2α mRNA expression level and clinicopathologic characteristics was found (Table 1), except for the pathological T stage (p = 0.026).

Association between AP-2α expression and RCC patient survival

We further analyzed the relationship between AP-2α mRNA expression and RCC patient prognosis. Kaplan‒Meier survival curves showed that higher AP-2α mRNA expression correlated with poor overall survival (p = 0.005) and disease-free survival (p = 0.005) (Fig. 3A, C). Similarly, patients with unmethylated AP-2α had poorer overall survival (p = 0.021) and disease-free survival (p = 0.028) compared to those with methylated AP-2α (Fig. 3B, D).

Fig. 3
figure 3

Kaplan‒meier survival curve based on AP-2α expression in patients with RCC. Kaplan‒meier analysis of survival probability (A, B) and disease-free survival (C, D) were performed based on AP-2α mRNA expression in RT‒qPCR assay (A, C) and COBRA assay (B, D). The -ΔCT value of relative AP-2α mRNA expression was calculated according to the following formula: -ΔCT = -(CTTarget – CTInternal). The median -ΔCT value was -12.8. Low expression was defined as -ΔCT ≤ -12.8, whereas high expression was defined as -ΔCT > -12.8 (p < 0.05 was considered to indicate significance)

Full size image

Univariate analysis revealed that higher AP-2α mRNA expression was associated with poor overall survival and disease-free survival [hazard ratio (HR): 5.101, 95% confidence interval (CI): 1.438–18.100, p = 0.012; HR: 5.160, 95% CI: 1.455–18.301, p = 0.011; respectively] (Table 2). After controlling for the clinical confounders that were statistically significant in the univariate analysis, multivariate analysis confirmed these findings, showing an increased risk for overall survival and disease-free survival (HR: 28.155, 95% CI: 3.798–208.698, p = 0.001; HR: 38.767, 95% CI: 4.822–311.671, p = 0.001; respectively) (Table 2).

Table 2 Univariate and multivariate analysis of risk factors on RCC patient overall survival and disease-free survival
Full size table

Considering the negative correlation between AP-2α mRNA expression and methylation status, we further explored the prognostic significance of AP-2α methylation. Multivariate analysis indicated a significantly increased risk for overall survival and disease-free survival in patients with unmethylated AP-2α (HR: 0.239, 95% CI: 0.065–0.871, p = 0.030; HR: 0.186, 95% CI: 0.041–0.832, p = 0.028; respectively) (Table 3).

Association between AP-2α expression and clear cell RCC patient survival

We investigated the relationship between AP-2α mRNA expression and clear cell RCC prognosis due to the significant differences among different histological types (Table 1). Kaplan‒Meier survival curves showed that patients with higher AP-2α mRNA expression had significantly poorer overall survival (p = 0.009) and disease-free survival (p = 0.007) (Fig. 4A, E). Similarly, patients with unmethylated AP-2α had poorer overall survival (p = 0.039) and disease-free survival (p = 0.043) than those with methylated AP-2α (Fig. 4B, F). Using the TCGA database, we confirmed that higher AP-2α mRNA expression was associated with poorer overall and disease-free survival in both clear cell (Fig. 4C, G) and papillary RCC (Fig. 4D, H) (overall survival, p < 0.001 and p < 0.001, respectively; disease-free survival, p < 0.001 and p < 0.01, respectively).

Fig. 4
figure 4

Kaplan‒Meier survival curve based on AP-2α expression in patients with different RCC subtypes. Kaplan‒Meier analysis of survival probability (A, B) and disease-free survival (E, F) were performed based on AP-2α mRNA expression in patients with clear cell RCC in RT‒qPCR assay (A, E) and COBRA assay (B, F). The -ΔCT value of relative AP-2α mRNA expression was calculated according to the following formula: -ΔCT = -(CTTarget – CTInternal). The median -ΔCT value was − 13.1. Low expression was defined as -ΔCT ≤ -13.1, whereas high expression was defined as -ΔCT > -13.1 (p < 0.05 was considered to indicate significance). The expression pattern of AP-2α in clear cell RCC (C, G) and papillary RCC (D, H) was predicted by the GEPIA2 database

Full size image

Univariate analysis identified AP-2α mRNA expression as a major risk factor for overall and disease-free survival in clear cell RCC (HR: 9.558, 95% CI: 1.195–76.484, p = 0.033; HR: 10.151, 95% CI: 1.268–81.246, p = 0.029, Table 4). Multivariate analysis further revealed that AP-2α mRNA expression was an independent risk factor for disease-free survival (HR: 21.421, 95% CI: 1.777–258.210, p = 0.016) but not overall survival in clear cell RCC (Table 4).

Table 3 Effect of AP-2α methylation status on RCC patient overall survival and disease-free survival
Full size table
Table 4 Univariate and multivariate analysis of risk factors on clear cell RCC patient overall survival and disease-free survival
Full size table

Discussion

The AP-2 family consists of different TFs that bind to the consensus DNA-binding sequence GCCNNNGGC [30, 31]. Among these, AP-2α, encoded by the TFAP2A gene, plays a significant role. Genetic and epigenetic alterations, such as DNA copy number variation and CpG island methylation, could influence gene expression during tumor pathogenesis [32, 33]. Our RT‒qPCR and COBRA assay results revealed that variable AP-2α mRNA expression across different RCC subtypes. We hypothesize that this deregulation in clear cell RCC is partly due to DNA methylation, as indicated by high methylation status of AP-2α detected by the COBRA assay. However, NimbleScan only predicted the methylation status of selective site, and not all methylation sites of AP-2α were examined. This suggests that other methylated sites may contribute to the observed variability in AP-2α mRNA expression in clear cell RCC.

To date, the exact role of AP-2α in RCC remains unclear. Different RCC subtypes are known to have distinct genetic bases and precursor cells [34, 35]. Previous studies, such as by Oya et al., demonstrated differential expression of AP-2 isoforms in RCC and normal adult kidneys, with clear cell RCC showing higher nuclear AP-2α expression, suggesting a functional role in this subtype [16]. However, their study found no significant difference in AP-2α expression based on tumor stage and grade. Our current results show that higher AP-2α mRNA expression is associated with high pathological T stage, poor overall survival, and poor disease-free survival in RCC (Table 1; Fig. 3). Both univariate and multivariate analyses indicated that AP-2α mRNA expression is a major risk factors for overall and disease-free survival in RCC patients (Table 2). We observed significant variability in AP-2α mRNA expression among RCC subtypes, with a similar impact on survival noted in clear cell RCC (Table 4; Fig. 4). We propose differing precursor cells may influence AP-2α function. Additionally, our results suggest that AP-2α functions differently in aggressive versus less aggressive tumors. Higher AP-2α mRNA expression levels were found in oncocytoma and chromophobe RCC compared to clear cell and papillary RCC. However, higher expression was significantly associated with poor survival in clear cell and papillary RCC patients (Fig. 4). We hypothesize that AP-2α plays an oncogenic role in aggressive RCC subtypes like clear cell RCC, but not in less aggressive subtypes. The inconsistency with Oya’s study might be due to different experimental approaches or populations used in assessing AP-2α expression in RCC [16]. AP-2 isoforms can form homo- or heterodimer complexes and bind differently to target DNA [14, 36], suggesting that their mechanisms and roles may differ in a tissue-specific manner.

Although AP-2α is known to play important roles in biological processes and is often associated with cancer progression and metastasis as a tumor suppressor gene [17,18,19], our results showed that higher AP-2α expression is linked to poor overall and disease-free survival in clear cell RCC. This suggests that AP-2α may play an oncogenic role in RCC, aligning with findings in other studies [20, 37, 38]. Studies have shown similar associations of AP-2α with poor prognosis in various cancers. For instance, Shi et al. found that AP-2α overexpression correlates with advanced tumor stage and poor prognosis in nasopharyngeal carcinoma, and it regulates cell growth and survival via the HIF-1α-mediated VEGF/PDGF pathway [20]. Yamashita et al. reported that AP-2α overexpression is linked to metastasis and recurrence in basal-squamous bladder cancer [37]. Liao et al. found that higher AP-2α expression predicts worsen prognosis in lung adenocarcinoma [38]. Furthermore, specific long non-coding RNAs or microRNAs (miRNAs) have been found to modulate AP-2α levels, influencing tumor biology. For example, Qin et al. identified 8 prognostic miRNAs in clear cell RCC and proposed a TF-miRNA signature involving AP-2α that could improve prognosis predictions [39]. Liang et al. discovered that Linc00467 enhances invasion and inhibit apoptosis in head and neck squamous cell carcinoma through the miRNA-1285-3p/AP-2α axis [40]. Yang et al. found that AP-2α acts as an oncogene in oral squamous cell carcinoma via the AP-2α-AS1/miRNA-1297/AP-2α axis [41]. Our study’s findings raise the possibility that AP-2α may function similarly in RCC, but further investigation is needed. Limitations of our study include the small number of cases for different RCC subtypes and the incomplete methylation status prediction by NimbleScan software, which restricted the verification of AP-2α methylation comprehensively.

In conclusion, our results showed that AP-2α was deregulated in RCC and that higher AP-2α expression was associated with poor overall and disease-free survival in RCC patients. These findings strongly suggest that AP-2α plays a crucial role in the pathophysiology of RCC. Further studies are needed to confirm its specific function in RCC and its potential as a prognostic marker.