CENPA expression elevated in PRCC samples

CENPA expression in 289 tumor tissues was substantially higher than that in 32 normal tissues (P < 0.001) (Fig. 1A), and it was also higher in 31 tumor tissues (P < 0.001) (Fig. 1B). Moreover, CENPA expression exhibited promising discrimination ability, with an AUC value of 0.936 in distinguishing tumors from normal tissues (Fig. 1C).

Fig. 1

CENPA expression between cancer and normal tissues in PRCC patients (A). CENPA expression levels in PRCC and matched normal tissues (B). ROC analysis of CENPA shows promising discrimination power between tumor and normal tissues (C)

Association between CENPA expression and clinicopathological characteristics

As revealed by Kruskal–Wallis rank-sum test, CENPA expression was related to pathological T stage (P < 0.001), pathological N stage (P < 0.001), and pathological M stage (P < 0.001), clinical stage (P < 0.001), and primary treatment effects (P = 0.007) (Fig. 2A–E). Furthermore, KM survival analysis demonstrated that high CENPA expression was strongly associated with poor prognosis (P < 0.001) compared with low CENPA expression (Fig. 3A, B).

Fig. 2

Association of CENPA expression with clinicopathologic characteristics. Pathological T stage (A); pathological N stage (B); pathological M stage (C); clinical stage (D) and primary therapy outcome (E)

Fig. 3

Kaplan–Meier survival curves comparing the high and low expression of CENPA in PRCC patients. Overall survival (OS) (A); Progression-free interval (PFS) (B)

Elevated CENPA expression predicts poor prognosis in different cancer stages

As revealed by multivariate survival analysis, PRCC cases in the following tumor stages were associated with dismal survival, including clinical stage (I & II, P = 0.037, III & IV, P < 0.001), pathological T stage (T1 & T2, P = 0.034, T3 & T4, P < 0.001), pathological N stage (N0, P = 0.054, N1 & N2, P = 0.003), and pathological M stage (M0, P = 0.383) (Fig. 4A). Therefore, CENPA level affected PRCC prognosis in diverse clinical stages. At the same time, in order to investigate whether the expression of CENPA helps to improve the judgment of pathological TNM and clinical staging, we also performed ROC analysis for each subgroup. The AUC results showed that the predictive accuracy of pathological TNM and clinical stage was significantly improved after the inclusion of CENPA expression (T stage: 0.603–0.936; N stage: 0.611–0.889;M stage: 0.630–0.897; clinical stage: 0.581–0.928) (Fig. 4B–I).

Fig. 4

Multivariate survival analysis of OS concerning CENPA expression in patients of different subgroups according to cancer stage (A). ROC analysis of predictive accuracy in different subgroups, including pathological T stage (B), pathological T stage with CENPA expression (C), pathological N stage (D), pathological N stage with CENPA expression (E), pathological M stage (F), pathological M stage with CENPA expression (G), clinical stage (H) and clinical stage with CENPA expression (I)

Role of CENPA in predicting PRCC survival under diverse clinicopathological characteristics

For further exploring CENPA-related mechanism within PRCC, this study assessed the relation of CENPA level with clinicopathological characteristics among PRCC cases through univariate Cox regression. More clinicopathological factors related to dismal prognosis included height, weight, primary treatment outcome, pathological T stage, and pathological N stage. For exploring survival-related factors, multivariate Cox regression was conducted on these factors. As a result, CENPA up-regulation was still a factor that independently predicted dismal OS, and other prognostic factors included height, weight, pathological T stage, pathological N stage and primary treatment outcome (Table 1).

Table 1 Association of clinicopathological characteristics with overall survival using univariate or multivariate Cox regression analysisConstruction and validation of the nomogram based on CENPA expression

In trial cohort, the nomogram was established by integrating survival-related clinicopathological characteristics identified from multivariate analysis (height, weight, pathological T stage, pathological N stage and CENPA expression level), so as to offer a quantitative approach for clinicians (Fig. 5A) (Due to the small sample size, primary treatment outcome was not included in the nomogram production finally). These variables were incorporated into the nomograms based on multivariable Cox analysis using a point scale. We then accumulated the position of the variable and denoted it as a total score. Thereafter, we drew vertical lines from total-point axis along outcome axis to determine the 1-, 3- and 5-year survival probabilities for PRCC cases. After calculation, our nomogram had a C-index value being 0.822 (95% CI:0.779–0.865). Moreover, as observed from calibration graph, the deviation alignment line approached the ideal curve both in trial and validation cohorts, which indicated that those predicted results were well consistent with actual results (Fig. 5B, C). We also calculated the C-index, which included only other prognostic indicators after removing CENPA expression. The result was lower than our nomogram, which was 0.804 (95% CI:0.757–0.851). This indicated that CENPA expression can assist TNM stage to predict OS of PRCC to a certain extent. On the whole, our nomogram was a superior approach for predicting long-time PRCC prognosis to diverse prognostic factors.

Fig. 5

Relationship between CENPA and other clinical factors with OS. Nomogram for predicting the probability of 1-, 3-, and 5-year OS for PRCC patients in trial cohort (A). Calibration plot of the nomogram for predicting the OS likelihood in trial cohort (B) and validation cohort (C)

Determination of CENPA expression in the samples

IHC staining was positive for the appearance of blue-purple particles. The score was rated in accordance with the percentage of positive cells and staining intensity. The scores of positive cell percentage were shown below, 0 (< 5%), 1 (5–25%), 2 (26–50%), 3 (51–75%), and 4 (> 75%), while the scores of staining intensity were as follows, 0 (no color or unclear), 1 (light purple), 2 (blue purple), and 3 (dark purple). Thereafter, the staining index (SI) was calculated as the product of positive cell percentage score and staining intensity score, which was expressed as negative for SI < 2 and positive for SI > 2. IHC results showed that the CENPA-positive rate was lower in control group than in PRCC group (Fig. 6A, B).

Fig. 6

Expression of CENPA in the two gruops detected by IHC staining (× 400) A:Control group; B:PRCC group)

Identification of DEGs between high and low CENPA expression groups

We utilized the R software DSEeq2 package to analyze TCGA-derived data (adjusted P < 0.05 and |logFC|> 1.5) and discovered altogether 1210 DEGs between CENPA up-regulation and down-regulation groups, among which, 1170 showed up-regulation while 40 showed down-regulation in CENPA up-regulation group (Fig. 7A, B).

Fig. 7

Differentially expressed genes between patients with high and low CENPA expression. Volcano plot of differentially expressed genes between the high and low CENPA expression groups. Normalized expression levels are shown in descending order from blue to red (A). Heatmap of the top ten significant differentially expressed genes between the high and low CENPA expression groups. Blue and red dots represent downregulated and upregulated genes, respectively (B)

Functional annotation and prediction of signaling pathways

For further understanding CENPA’s role in PRCC prognosis based on those 1210 DEGs discovered between CENPA up-regulation and down-regulation groups, we carried out GO annotation. Finally, we obtained 233 GO-biological process (GO-BP) terms, such as nuclear division, chromosome segregation, and mitotic nuclear division (Fig. 8A). The above findings indicated that abnormal CENPA level was related to nuclear division. In the meantime, we also discovered 18 GO-cellular component (GO-CC) terms, which confirmed that the abnormal CENPA expression was associated with kinetochore (Fig. 8B). In addition, the GO-molecular function (GO-MF) terms identified included the DNA-binding transcription activator activity, RNA polymerase II-specific was significantly enriched (Fig. 8C).

Fig. 8

Enriched GO terms in the “biological process” category (A).Enriched GO terms in the “cellular component” category (B). Enriched GO terms in the “molecular function” category (C). Blue and red tones represent adjusted P values form 0.0–0.05, respectively, and different circle sizes represent the number of DEGs

The CENPA-associated signaling pathway based on GSEA

Subsequently, signaling pathways related to PRCC were identified between high and low CENPA expression groups through GSEA using significant difference in MSigDB Collection (c2.all.v7.0) (adjusted P < 0.05, FDR < 0.25). Finally, 5 pathways, including neuroactive ligand receptor interactions, cytokine receptor interactions, extracellular matrix regulators, extracellular matrix glycoproteins and nuclear matrisome, were identified with significant differences between both groups (Fig. 9A–E).

Fig. 9

Enrichment plots from GESA. Several pathway were differentially enriched in PRCC patients according to high and low CENPA expression. Neuroactive ligand receptor interactions (A), cytokine receptor interactions (B), extracellular matrix regulators (C), extracellular matrix glycoproteins (D), nuclear matrisome (E). (ES, enrichment score; NES, normalized enrichment score; ADJ p-Val, adjusted P-value; FDR, false discovery rate).

Correlation between CENPA expression and immune infiltration

At last, this study examined the relation of CENPA level (TPM) with the enrichment levels of immune cells (obtained from ssGSEA) according to Spearman correlation analysis. As a result, CENPA level showed negative relation to enrichment levels of B cells, Cytotoxic cells, Nrutrophlis, NK CD56bright cells, CD8 T cells, TEM, iDC, Macrophages, DC, Mast cells, and Eosinophils. Moreover, CENPA level was positively associated with enrichment levels of Tcm, T cells, NK cells, Th17 cells, TReg, Th1 cells, TFH, T CD56dim cells, pDC, aDC, Th2 cells and Tgd (Fig. 10).

Fig. 10

Correlations between the relative abundance of 24 immune cells and CENPA expression levels. The size of the dots represents the absolute Spearman’s correlation coefficient values