Background: To improve the clinical evaluation of the prognosis of papillary renal cell carcinoma (PRCC), we screened a model to predict the survival of patients with mutations in related genes. Methods: We downloaded RNA sequencing information from all patients with PRCC in TCGA. We first analyzed the differences in genes and the enrichment of these differences. Then, by selecting mutant genes, constructing a protein-protein interaction network, least absolute shrinkage and selection operator regression, and multivariable Cox regression, a prognosis model was constructed. Additionally, the model was validated using external data sets. We analyzed the immune infiltration of PRCC and the correlation between the model and popular targets. Finally, we performed tissue microarray analysis and immunohistochemistry to verify the expression levels of the three genes. Results: We constructed a three-gene (never in mitosis gene A-related kinase 2 [NEK2], centromere protein A [CENPA], and GINS complex subunit 2 [GINS2]) model. The verification results indicated that the model had a good prediction effect. We also developed a visual nomogram. Enrichment analysis revealed the major pathways involved in muscle system processes. Immunoassays showed that the expression level of CENPA was positively correlated with PD-1 and CTLA4 expression levels. Immunohistochemical and tissue microarray results showed that these three genes were highly expressed in PRCC, which was consistent with the predicted results in the database. Conclusion: We constructed and verified a three-gene model to predict the patient survival. The results show that the model has a good prediction effect.

Renal cell carcinoma (RCC) is one of the most common cancers and mainly affects men. Papillary RCC (PRCC) is the second most common renal cancer, accounting for 10% of all renal cancers [1]. The typical pathological feature of PRCC is the papillary process, and its pathological features are usually divided into types 1 and 2 [2, 3]. Type 1 is characterized by a single layer of small cells with small quantity of cytoplasm, whereas type 2 is dominated by eosinophils, often with a higher number of cells and worse outcomes [4].

In recent years, the treatment of related mutation sites has gradually become a research direction in RCC. Studies have shown that MET mutations are associated with type 1 PRCC [5]. Therefore, therapeutic drugs for this gene mutation have become a research hotspot, including crizotinib [6], savolitinib [7], and cabozantinib [8, 9]. Type 2 PRCC is associated with changes in expression levels of CDKN2A, as well as of the chromosome-modifying genes SETD2, BAP1, and PBRM1 [10]. In addition, TFE3 translocation is related to the prognosis of RCC [11].

However, the prognosis of PRCC is poor. We analyzed the TCGA database and constructed a mutation-related prognosis model based on known survival information. This model can predict the survival of patients and provide a reference for clinicians. Furthermore, PD-1 and CTLA4 inhibitors have recently been used widely in patients with RCC [12, 13]. Therefore, we performed immunoassay for some popular drug targets.

Data Collection and Data Preprocessing

RNA sequencing (RNA-seq) data were downloaded from the UCSC Xena official website (https://xena.ucsc.edu/), which includes 321 cases of RNA-seq expression data (32 normal samples and 289 tumor samples) and their corresponding clinical and survival data. We matched the clinical information with the survival data and removed the samples not in the expression data. In addition, we removed the samples with incomplete survival information or follow-up duration of <30 days. Finally, we removed samples with only survival information and no expression data. Overall, 276 tumor samples were used for survival analysis. These survival data were randomly divided into a 1:1 group using the “createdatapartition” function of the “caret” package: 138 cases in the training group and 138 cases in the test group. In addition, we downloaded the “mutect” mutation data of 281 patients on the UCSC Xena, which obtained the mutation information of 11,610 genes.

Application of the Protein-Protein Interaction Network and Construction of the Model

In R language, 321 samples were analyzed for gene differences after voom standardization with the “limma” package. Genes with mutation information and differential genes were intersected to obtain differentially expressed genes (DEGs) with mutation significance. After removing the genes, they were uploaded to the STRING (https://cn.string-db.org) to construct a protein-protein interaction network and then imported into the Cytoscape software. The MCODE plug-in was used to select the sub-network with the highest score. In the training group, we further screened genes related to prognosis using the least absolute shrinkage and selection operator (LASSO) model. We constructed a multifactor COX regression model using a stepwise regression method. Akaike information criterion (AIC) was used to evaluate the model. Finally, we screened for the three-gene model with the lowest AIC value.

Establishment of the Risk Score and Construction of a Nomogram

According to this model, we established a risk score to predict prognosis. Risk score = (coef gene1 * exp gene1) + (coef gene2 * exp gene2) + (coef gene3 * exp gene3). To evaluate the prognosis of patients at 1, 3, and 5 years, we used the risk score, age, gender, and stage to establish the nomogram model in the training set and draw its calibration curves at 1, 3, and 5 years, respectively.

Validation of the Test Group and Analysis of the Immune Microenvironment

We used the optimal cutoff value to divide the test groups to verify the model. Based on these values, a survival curve was constructed for the test group. Then the “predict” function was used to calculate the risk score corresponding to each sample, according to the formula: risk score = (0.9311) * (expression level of never in mitosis gene A-related kinase 2 [NEK2]) + (1.3564) * (expression level of centromere protein A [CENPA]) + (−1.0755) * (expression level of GINS complex subunit 2 [GINS2]). We used the “median” function to calculate the median of risk scores in R studio. Finally, all tumor samples were divided into high-risk and low-risk groups for immunoassay. The overall immune microenvironment was compared with the single-sample gene set enrichment analysis [14] of the “GSVA” package, and the correlation of specific popular immune targets was analyzed.

Functional and Pathway Enrichment Analyses

The screened DEG, including 705 upregulated genes and 1,702 downregulated genes, were enriched by the “clusterProfiler” package for gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis. Potential relationships between genes and pathways were studied using enrichment analysis.

Patient Samples

We extracted all pRCC cases from the Ruijin Hospital affiliated with Shanghai Jiaotong University Medical College between 2020 and 2021, including a total of 21 cases. All samples were diagnosed by two pathologists. All patients provided their written informed consent before their enrolment. This study protocol was reviewed and approved by the Ruijin Hospital Ethics Committee, Shanghai Jiao Tong University School of Medicine (approval number: 2022-099). The clinical pathological information of all patients was presented in the supplementary materials (shown in online Supplementary S2; for all online suppl. material, see https://doi.org/10.1159/000539096).

Immunohistochemistry

Paraffin tissue chip 4-μM serial sections were pasted on slides, placed in a 75°C oven for 4–5 h, and maintained at 95°C constant temperature for 20 min. The primary antibody was diluted to 1:100 and stored at 4°C. Automated immunohistochemistry was performed using the Autostainer 48 Link instrument (Dako, Glostrup, Denmark) with appropriate positive controls. We used EnVision FLEX+, Mouse, High pH (Dako, CAT#K8002, Glostrup, Denmark), according to the manufacturer’s recommendations. All immunohistochemical results were evaluated jointly by two experienced pathologists. The following antibodies were used: NEK2 (H00004751-M01; Novus, 1:100), CENPA (NBP3-13389; Novus, 1:500), and GINS2 (NBP2-33825; Novus, 1:500).

Preparation of Tissue Microarrays and Semiquantitative Score

First, the paraffin tissue was embedded in a Quick-Ray Recipient Block (UniTMA, UB06-2, Korea). The samples were placed in an oven at 70°C for 30–60 min. After the wax block was removed from the oven in a transparent state, the tissue microarray (TMA) block was placed in the cassette embedding box, and the appropriate boiling liquid white wax was added for embedding. We divided the immunohistochemical results into three cases for semiquantitative scoring: 1 point: <10% of the proteins expressed in tumor cells were considered negative (−); 2 points: 10–50% of tumor cells were considered to be focal or weakly expressed (−/+); 3 points: >50% strong expression and diffusion were considered positive (+). Positive expression of NEK2, CENPA, and GINS2 were defined as their positive nuclear reactivity in the neoplastic cells. Cytoplasmic and membrane expressions were not considered staining-positive. All results were scored separately by two pathologists.

Statistical Analysis

All analyses were performed using the R software (4.1.2). Except for special instructions, all the other results were statistically different according to the standard of p < 0.05.

Mutation-Related DEGs in PRCC

All workflows were displayed (shown in Fig. 1). In the 321 samples of gene expression data, DEGs meeting the requirements were screened in the gene expression data according to the standard of |log2foldchange| > 2 and p < 0.05. At this level, we identified 705 upregulated genes and 1,702 downregulated genes; these genes were presented in the form of a volcano plot (shown in Fig. 2a). These genes were crossed with the mutation information in TCGA, and 1,082 differential genes with mutation significance were obtained. By mapping the mutation spectrum of the top 30 genes, we found that most were missense mutations in PRCC. Among them, the proportion of TTN, MUC16, and KMT2C was 8%, which may be related to the prognosis of PRCC.

Fig. 1.

Flowchart of the whole work.

Fig. 1.

Flowchart of the whole work.

Close modal
Fig. 2.

a Volcano plot of differential genes. b, c LASSO regression. d PPI network from Cytoscape. PPI, protein-protein interaction.

Fig. 2.

a Volcano plot of differential genes. b, c LASSO regression. d PPI network from Cytoscape. PPI, protein-protein interaction.

Close modal

Screening of Genes and Description of Each Gene

The relationship between each DEG was shown in the protein-protein interaction network on the STRING website. We downloaded the interaction data of the web page and uploaded them to the Cytoscape software to obtain a graph, which included 1,002 nodes and 5,616 edges. According to the standards of degree cutoff = 2, node score cutoff = 0.2, k-core = 2, and max.depth = 100, the sub-network with the highest score (score: 27.484) was selected for MCODE, which consisted of 32 genes (shown in Fig. 2d). We selected the appropriate lambda value in the training group and screened the genes with a coefficient of not 0 by LASSO regression (shown in Fig. 2b, c). We used stepwise regression to select the lowest AIC value for multivariate Cox regression and finally obtained a three-gene formula related to prognosis: risk score = (0.9311) * (expression level of NEK2) + (1.3564) * (expression level of CENPA) + (−1.0755) * (expression level of GINS2). Among them, NEK2 and CENPA expression levels were positively correlated with the patient prognosis, whereas the GINS2 expression level was negatively correlated. The contents of these three genes in normal and tumor samples were displayed in a violin diagram that showed that the levels of these three genes varied significantly between the two groups of samples (shown in Fig. 3a–c). A survival curve for each gene was also constructed, which was significantly correlated with the prognosis (shown in Fig. 3d–f). Furthermore, we analyzed gene expression, mutations, and drug resistance from Broad-Novartis Cancer Cell Line Encyclopedia (CCLE) and the COSMIC Cell Lines Project (CCLP) databases (shown in online Supplementary S3).

Fig. 3.

a–c Violin plot of three gene expression in normal and tumor tissues. d–f Survival curves of three genes.

Fig. 3.

a–c Violin plot of three gene expression in normal and tumor tissues. d–f Survival curves of three genes.

Close modal

Construction of a Nomogram

Based on the clinical information of patients, we predicted the survival of the patients based on the four variables of age, gender, stage, and risk score. Thus, we plotted nomograms to show the total scores at 1, 3, and 5 years (shown in Fig. 4a) and their respective calibration curves (shown in Fig. 4b–d). In general, the model had a good predictive effect. This could help clinicians judge the prognosis of patients and improve the treatment effectiveness.

Fig. 4.

a–d Nomogram and corresponding calibration curves for 1, 3, and 5 years.

Fig. 4.

a–d Nomogram and corresponding calibration curves for 1, 3, and 5 years.

Close modal

Verification of the Model

Next, we verified the model using a test group. First, we used the “predict” function to predict the test group values with the three-gene model. The results showed that the C-index was 0.8046, indicating that the prediction effect of the model was significant. We selected the optimal cutoff point and divided the values into high-risk and low-risk groups to draw a survival curve (shown in Fig. 5b). We observed that the survival of the high-risk group was significantly shorter than that of the low-risk group, and the p value was statistically significant. Subsequently, we plotted the timeROC curves for 1, 3, and 5 years and calculated their AUC values (shown in Fig. 5c). The prediction effect of the model was good, particularly for 3 years.

Fig. 5.

Verification of the testing group. a Multivariate Cox regression with other clinical information. b Survival curves of the risk score at different levels. c ROC curves and corresponding AUC values for 1, 3, and 5 years. d Expression of three genes in high-risk and low-risk groups. e Comparison of the scatter diagram and heat map of different levels of the risk score.

Fig. 5.

Verification of the testing group. a Multivariate Cox regression with other clinical information. b Survival curves of the risk score at different levels. c ROC curves and corresponding AUC values for 1, 3, and 5 years. d Expression of three genes in high-risk and low-risk groups. e Comparison of the scatter diagram and heat map of different levels of the risk score.

Close modal

In addition, we grouped all samples according to the optimal cutoff point and ranked them according to the risk score from lowest to highest (shown in Fig. 5e). A time-sample scatter plot was drawn based on this information. The survival time of the high-risk group was significantly shorter than that of the low-risk group. Heat maps were generated for the three genes in the model. We found that NEK2 and CENPA had lower expression levels in the low-risk group than in the high-risk group. To verify these results, we also plotted the expression levels in a box plot, which was consistent with the above conclusion (shown in Fig. 5d).

Finally, the risk score and three clinical variables (age, gender, and stage) were used for multivariate Cox regression (shown in Fig. 5a). We obtained a C-index of 0.9 for the model, which showed a splendid prediction effect and statistically significant results.

Functional and Pathway Enrichment Analyses by GO and KEGG

Using the package “clusterProfiler” for GO analysis in R language, the DEGs were divided into three parts for enrichment analysis: molecular function, biological process, and cellular components. The muscle system and muscle contraction were enriched in biological processes. In addition, DEGs were enriched in the cellular components of the apical part of the cell and collagen-containing extracellular matrix. In terms of the molecular function, DEGs were primarily enriched in metal ion transmembrane transporter activity and passive transmembrane transporter activity (shown in Fig. 6a, b). Finally, in the network diagram of the interaction between pathways, it was shown that calcium ion transport, regulation of ion transport, potassium ion transport, and heart contraction were the main intersections of each pathway (shown in Fig. 6d, e). Moreover, the downregulated DEGs were significantly enriched in the calcium signaling pathway, followed by the neuroactive ligand receiver interaction in the KEGG analysis (shown in Fig. 6c). These pathways may be associated with the prognosis and survival of patients with PRCC.

Fig. 6.

a, b Enrichment results of DEGs in GO. c Enrichment results of DEGs in KEGG. d, e Network between pathways and genes, pathways and pathways.

Fig. 6.

a, b Enrichment results of DEGs in GO. c Enrichment results of DEGs in KEGG. d, e Network between pathways and genes, pathways and pathways.

Close modal

Correlation Immunoassay

First, we used the ESITIMATE algorithm [15] to calculate the stromal and immune scores of normal and tumor samples of PRCC, respectively (shown in Fig. 7e). We observed that the degree of immune cell infiltration in tumor samples was significantly higher than that in normal tissues, while stromal cells have a slightly lower number than those in normal tissues. Thereafter, we divided the samples into high-risk and low-risk groups according to the median risk score. According to the data collected by previous studies [16], we used the single-sample gene set enrichment analysis method to calculate the immune infiltration of 28 immune cells in the two groups (shown in Fig. 7a). The box diagram shows that the expression of each immune cell in the high-risk group was slightly higher than that in the low-risk group. Next, in order to further explore the relationships between key genes and tumor immune infiltration, we performed correlation analysis. We mapped the correlations between the three genes and 28 immune cell types (shown in Fig. 7b), which showed that the three genes were positively correlated with the levels of activated CD4 T cells and type 2 helper cells and negatively correlated with the levels of immature dendritic cells. The calculated p values indicated that the abovementioned differences were statistically significant. Finally, we selected several common immune loci and plotted their correlation with the expression of the three genes (shown in Fig. 7c). The figure showed that the CENPA expression level was strongly correlated with CTLA4 and PD-1 expression levels. We plotted the correlation between these two factors (shown in Fig. 7d).

Fig. 7.

a Box plot of immune cell infiltration between different levels of the risk score. b Correlation between three genes and different immune cells. c Correlation between three genes and several popular immune cells. d Graph of correlation between CENPA and CTLA4, PD-1. e Comparison of ESTIMATE scores between tumor samples and normal samples.

Fig. 7.

a Box plot of immune cell infiltration between different levels of the risk score. b Correlation between three genes and different immune cells. c Correlation between three genes and several popular immune cells. d Graph of correlation between CENPA and CTLA4, PD-1. e Comparison of ESTIMATE scores between tumor samples and normal samples.

Close modal

Validation of Immunohistochemistry Results

As regards to the expression of the NEK2 gene, we first verified the expression of the NEK2 antibody in testicular tissue, which stained the nucleus. Therefore, we believe that the testicular tissue served as a positive control (shown in online Supplementary S1). However, in PRCC and adjacent tissues, almost all cells were stained in the cytoplasm and cell membrane. Interestingly, in normal renal tissue, this gene was mainly expressed in the glomerulus and distal convoluted tubules. In PRCC, there was diffuse staining (shown in Fig. 8a). This may indicate that PRCC initially occurs in the distal convoluted tubules.

Fig. 8.

a Immunohistochemical results for three genes. b Immunohistochemical results of TMA for three genes. c Semiquantitative score statistics for three genes. TMA, tissue microarray.

Fig. 8.

a Immunohistochemical results for three genes. b Immunohistochemical results of TMA for three genes. c Semiquantitative score statistics for three genes. TMA, tissue microarray.

Close modal

CENPA is expressed in the nucleus of PRCC and normal tissues, which is consistent with the expected results of antibody staining. In addition, CENPA had higher expression levels in tumor tissue than in normal tissue (shown in Fig. 8a).

In addition, the expression of GINS2 was altered in normal and tumor tissues. An antibody was used for nuclear staining, so our standard was to consider nuclear staining as a positive result. The results showed that GINS2 was scattered and expressed in the tumor tissue. However, in normal tissues, GINS2 staining was almost complete in the cytoplasm and cell membrane, so we considered it to be a negative result (shown in Fig. 8a).

Finally, to evaluate the widespread distribution of these three genes in PRCC tissues, we created a TMA from paraffin-embedded tissues of patients with PRCC in Ruijin Hospital from 2020 to 2022. The corresponding immunohistochemistry of the three genes was performed (shown in Fig. 8b). The expression levels of these three genes were higher in tumor tissue than those in normal tissues. The difference was statistically significant (shown in Fig. 8c). The results are consistent with the analysis of the TCGA database.

Gene mutations are often a cause of cancer. In recent years, an increasing number of studies have explored new mutation sites and screened for new tumor markers. Most of these genes are used for diagnosis and have corresponding target therapies. In this study, we used bioinformatics and statistical methods to analyze RNA-seq and survival information of PRCC. The mutation information was obtained from the TCGA database, and the DEGs with mutation information were selected. We conducted LASSO regression, followed by multivariate Cox regression. Finally, the lowest AIC value was selected to construct a three-gene risk score. A nomogram was used to directly reflect the survival of patients at 1, 3, and 5 years. To prove the practical usefulness of the model, we used an external validation set to predict the model results. After grouping the samples, the survival curves of high-risk and low-risk patients were drawn. We also plotted the ROC curves at 1, 3, and 5 years and calculated the corresponding AUC values. The validation results showed that the model had good accuracy and prognostic value. Finally, we created a TMA from paraffin-embedded PRCC tissues and performed immunohistochemistry. The results were scored semiquantitatively by two pathologists. We found that the protein expression level of these three genes were consistent with those in the database.

NEK2 is a mitotic cyclin, and changes in its expression level can lead to various diseases [17]. Studies have reported that NEK2 may be associated with the occurrence of polycystic kidney diseases [18]. NEK2 is also an important target for many cancers [19]. High expression levels of NEK2 significantly affect the survival of patients with renal clear cell carcinoma [20] and may be associated with the Wnt/β-catenin pathway [21]. Moreover, NEK2 may be related to cancer immunotherapy [19]. Recently, studies have found that a high expression level of NEK2 is closely related to the hypermethylation of CpG islands in the distal part of its promoter, which affects drug resistance and prognosis of multiple myeloma [22].

CENPA is a variant of histone H3 located in the centromeric region. It is a key epigenetic factor in the establishment and function of centromeric regions. Dynamic phosphorylation of CENPA ser68 may be involved in the assembly of CENPA in the centromeric region [23]. Studies have shown that CENPA is highly expressed in PRCC and leads to a poor prognosis, which is consistent with our results [24]. In other cancers, experiments have shown that CENPA is activated by β-catenin to promote the occurrence and metastasis of renal clear cell carcinoma [25] and may affect the prognosis of colon cancer by activating downstream KPNA [26].

GINS2 is a cyclic nucleic acid replication factor, which is essentially a replication helicase. GINS2 influences the development of pancreatic cancer by activating the ERK/MAPK signaling pathway [27]. It also affects the prognosis and metastasis of lung cancer [28], breast cancer [29], glioma [30], hepatocellular carcinoma [31], and ovarian cancer [32]. However, there is no previous evidence about the relationship between GINS2 and PRCC.

Furthermore, we conducted an enrichment analysis of DEGs. The results showed that the DEGs were mainly related to the muscle system process, metal ion transmembrane transporter activity, and calcium signaling pathway. Finally, we analyzed the immune infiltration of PRCC cells. We found that the CENPA expression level was positively correlated with several common immune targets, such as PD-1 and CTLA4. Immunotherapy is the main focus in the treatment of RCC [33], and this may also be the research direction for future studies.

We constructed an independent prognostic model of PRCC through analysis and sorting of the TCGA database. We also analyzed the protein levels of these three genes by immunohistochemistry. This three-gene scoring model can reliably predict the prognosis and therapeutic significance of patients with PRCC.

All patients provided their written informed consent before their enrolment. This study protocol was reviewed and approved by the Ruijin Hospital Ethics Committee, Shanghai Jiao Tong University School of Medicine (approval number: 2022-099).

The authors have no conflicts of interest to declare.

This study was supported by the National Natural Science Foundation of China (82002667) and the National Natural Science Foundation of China (81903050).

Xiangyun Li was responsible for writing the manuscript, Yang Liu conducted data analysis, Luting Zhou provided financial support, Jianhua Wang modified the manuscript, and Xiaoqun Yang provided scientific research ideas. All the authors reviewed the manuscript.

Additional Information

Xiangyun Li, Yang Liu, and Luting Zhou contributed equally to this work.

The data that support the findings of this study are not publicly available due to privacy protection and intellectual property considerations but are available from the corresponding author upon reasonable request.

1.
Mendhiratta
N
,
Muraki
P
,
Sisk
AE
Jr
,
Shuch
B
.
Papillary renal cell carcinoma: review
.
Urol Oncol
.
2021
;
39
(
6
):
327
37
.
2.
MacLennan
GT
,
Cheng
L
.
Five decades of urologic pathology: the accelerating expansion of knowledge in renal cell neoplasia
.
Hum Pathol
.
2020
;
95
:
24
45
.
3.
Linehan
WM
,
Ricketts
CJ
.
The Cancer Genome Atlas of renal cell carcinoma: findings and clinical implications
.
Nat Rev Urol
.
2019
;
16
(
9
):
539
52
.
4.
Courthod
G
,
Tucci
M
,
Di Maio
M
,
Scagliotti
GV
.
Papillary renal cell carcinoma: a review of the current therapeutic landscape
.
Crit Rev Oncol Hematol
.
2015
;
96
(
1
):
100
12
.
5.
Rhoades Smith
KE
,
Bilen
MA
.
A review of papillary renal cell carcinoma and MET inhibitors
.
Kidney Cancer
.
2019
;
3
(
3
):
151
61
.
6.
Schoffski
P
,
Wozniak
A
,
Escudier
B
,
Rutkowski
P
,
Anthoney
A
,
Bauer
S
, et al
.
Crizotinib achieves long-lasting disease control in advanced papillary renal-cell carcinoma type 1 patients with MET mutations or amplification. EORTC 90101 CREATE trial
.
Eur J Cancer
.
2017
;
87
:
147
63
.
7.
Choueiri
TK
,
Plimack
E
,
Arkenau
HT
,
Jonasch
E
,
Heng
DYC
,
Powles
T
, et al
.
Biomarker-based phase II trial of savolitinib in patients with advanced papillary renal cell cancer
.
J Clin Oncol
.
2017
;
35
(
26
):
2993
3001
.
8.
D’Angelo
A
,
Bagby
S
,
Di Pierro
G
,
Chirra
M
,
Nobili
S
,
Mini
E
, et al
.
An overview of the clinical use of cabozantinib in the treatment of advanced non-clear-cell renal cell carcinoma (NCCRCC)
.
Crit Rev Oncol Hematol
.
2020
;
149
:
102921
.
9.
Borchiellini
D
,
Barthélémy
P
.
Cabozantinib: a new first-line option for papillary renal cell carcinoma
.
Lancet
.
2021
;
397
(
10275
):
645
7
.
10.
The Cancer Genome Atlas Research Network
,
Linehan
WM
,
Spellman
PT
.
Comprehensive molecular characterization of papillary renal-cell carcinoma
.
N Engl J Med
.
2016
;
374
(
2
):
135
45
.
11.
Sun
G
,
Chen
J
,
Liang
J
,
Yin
X
,
Zhang
M
,
Yao
J
, et al
.
Integrated exome and RNA sequencing of TFE3-translocation renal cell carcinoma
.
Nat Commun
.
2021
;
12
(
1
):
5262
.
12.
Maughan
BL
.
Start of a new era: management of non-clear cell renal cell carcinoma in 2022
.
Curr Oncol Rep
.
2022
;
24
(
9
):
1201
8
.
13.
Rotte
A
.
Combination of CTLA-4 and PD-1 blockers for treatment of cancer
.
J Exp Clin Cancer Res
.
2019
;
38
(
1
):
255
.
14.
Barbie
DA
,
Tamayo
P
,
Boehm
JS
,
Kim
SY
,
Moody
SE
,
Dunn
IF
, et al
.
Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1
.
Nature
.
2009
;
462
(
7269
):
108
12
.
15.
Yoshihara
K
,
Shahmoradgoli
M
,
Martinez
E
,
Vegesna
R
,
Kim
H
,
Torres-Garcia
W
, et al
.
Inferring tumour purity and stromal and immune cell admixture from expression data
.
Nat Commun
.
2013
;
4
:
2612
.
16.
Charoentong
P
,
Finotello
F
,
Angelova
M
,
Mayer
C
,
Efremova
M
,
Rieder
D
, et al
.
Pan-cancer immunogenomic analyses reveal genotype-immunophenotype relationships and predictors of response to checkpoint blockade
.
Cell Rep
.
2017
;
18
(
1
):
248
62
.
17.
Kokuryo
T
,
Yokoyama
Y
,
Yamaguchi
J
,
Tsunoda
N
,
Ebata
T
,
Nagino
M
.
NEK2 is an effective target for cancer therapy with potential to induce regression of multiple human malignancies
.
Anticancer Res
.
2019
;
39
(
5
):
2251
8
.
18.
Dana
D
,
Das
T
,
Choi
A
,
Bhuiyan
AI
,
Das
TK
,
Talele
TT
, et al
.
Nek2 kinase signaling in malaria, bone, immune and kidney disorders to metastatic cancers and drug resistance: progress on Nek2 inhibitor development
.
Molecules
.
2022
;
27
(
2
):
347
.
19.
Huang
X
,
Zhang
G
,
Tang
T
,
Gao
X
,
Liang
T
.
One shoot, three birds: targeting NEK2 orchestrates chemoradiotherapy, targeted therapy, and immunotherapy in cancer treatment
.
Biochim Biophys Acta Rev Cancer
.
2022
;
1877
(
2
):
188696
.
20.
Wang
C
,
Huang
Y
,
Ma
X
,
Wang
B
,
Zhang
X
.
Overexpression of NEK2 is correlated with poor prognosis in human clear cell renal cell carcinoma
.
Int J Immunopathol Pharmacol
.
2021
;
35
:
20587384211065893
.
21.
Zhou
J
,
Lai
J
,
Cheng
Y
,
Qu
W
.
NEK2 serves as a novel biomarker and enhances the tumorigenicity of clear-CellRenal-cell carcinoma by activating WNT/β-Catenin pathway
.
Evid Based Complement Alternat Med
.
2022
;
2022
:
1890823
.
22.
Feng
X
,
Guo
J
,
An
G
,
Wu
Y
,
Liu
Z
,
Meng
B
, et al
.
Genetic aberrations and interaction of NEK2 and TP53 accelerate aggressiveness of multiple myeloma
.
Adv Sci
.
2022
;
9
(
9
):
e2104491
.
23.
Yu
Z
,
Zhou
X
,
Wang
W
,
Deng
W
,
Fang
J
,
Hu
H
, et al
.
Dynamic phosphorylation of CENP-A at Ser68 orchestrates its cell-cycle-dependent deposition at centromeres
.
Dev Cell
.
2015
;
32
(
1
):
68
81
.
24.
Li
J
,
Li
Q
,
Yuan
Y
,
Xie
Y
,
Zhang
Y
,
Zhang
R
.
High CENPA expression in papillary renal cell carcinoma tissues is associated with poor prognosis
.
BMC Urol
.
2022
;
22
(
1
):
157
.
25.
Wang
Q
,
Xu
J
,
Xiong
Z
,
Xu
T
,
Liu
J
,
Liu
Y
, et al
.
CENPA promotes clear cell renal cell carcinoma progression and metastasis via Wnt/β-catenin signaling pathway
.
J Transl Med
.
2021
;
19
(
1
):
417
.
26.
Liang
YC
,
Su
Q
,
Liu
YJ
,
Xiao
H
,
Yin
HZ
.
Centromere protein A (CENPA) regulates metabolic reprogramming in the colon cancer cells by transcriptionally activating karyopherin Subunit Alpha 2 (KPNA2)
.
Am J Pathol
.
2021
;
191
(
12
):
2117
32
.
27.
Huang
L
,
Chen
S
,
Fan
H
,
Ji
D
,
Chen
C
,
Sheng
W
.
GINS2 promotes EMT in pancreatic cancer via specifically stimulating ERK/MAPK signaling
.
Cancer Gene Ther
.
2021
;
28
(
7–8
):
839
49
.
28.
Sun
D
,
Zong
Y
,
Cheng
J
,
Li
Z
,
Xing
L
,
Yu
J
.
GINS2 attenuates the development of lung cancer by inhibiting the STAT signaling pathway
.
J Cancer
.
2021
;
12
(
1
):
99
110
.
29.
Peng
L
,
Song
Z
,
Chen
D
,
Linghu
R
,
Wang
Y
,
Zhang
X
, et al
.
GINS2 regulates matrix metallopeptidase 9 expression and cancer stem cell property in human triple negative Breast cancer
.
Biomed Pharmacother
.
2016
;
84
:
1568
74
.
30.
Shen
YL
,
Li
HZ
,
Hu
YW
,
Zheng
L
,
Wang
Q
.
Loss of GINS2 inhibits cell proliferation and tumorigenesis in human gliomas
.
CNS Neurosci Ther
.
2019
;
25
(
2
):
273
87
.
31.
Li
Z
,
Song
G
,
Guo
D
,
Zhou
Z
,
Qiu
C
,
Xiao
C
, et al
.
Identification of GINS2 prognostic potential and involvement in immune cell infiltration in hepatocellular carcinoma
.
J Cancer
.
2022
;
13
(
2
):
610
22
.
32.
Yan
T
,
Liang
W
,
Jiang
E
,
Ye
A
,
Wu
Q
,
Xi
M
.
GINS2 regulates cell proliferation and apoptosis in human epithelial ovarian cancer
.
Oncol Lett
.
2018
;
16
(
2
):
2591
8
.
33.
Braun
DA
,
Bakouny
Z
,
Hirsch
L
,
Flippot
R
,
Van Allen
EM
,
Wu
CJ
, et al
.
Beyond conventional immune-checkpoint inhibition: novel immunotherapies for renal cell carcinoma
.
Nat Rev Clin Oncol
.
2021
;
18
(
4
):
199
214
.