3.1 PTCD1 upregulation in ccRCC

Initially, researchers did a transcriptome differential analysis and discovered that PTCD1 expression level for ccRCC samples was considerably greater than that in normal samples (p = 1.053e−06) (Fig. 2a). According to paired analysis, PTCD1 transcription in ccRCC samples was substantially greater than that in adjacent tissues (p < 0.05) (Fig. 2b). A stratified analysis discovered that PTCD1 transcription steadily rose with stage (p = 9.435e−05), grade (p = 4.671e−04), T stage (p = 0.002), N stage (p = 0.006), and M stage (p = 1.102e-04), while age (p = 0.631) or gender (p = 0.71) seemed to have no relationship with PTCD1 transcription level (Fig. 2c–i).

Fig. 2

The PTCD1 expression level and survival analysis in ccRCC. a PTCD1 expression in normal tissues and ccRCC tissues; b PTCD1 expression in ccRCC paired tissues; Box plot evaluating PTCD1 expression according to different clinical characteristics including age (c), gender (d), stage (e), grade (f), T stage (g), N stage (h) and M stage (i); The OS (j), DSS (k), PFI (l) of high-PTCD1 versus low-PTCD1 in the TCGA dataset; ROC curve demonstrated the stability of PTCD1 in evaluating 1-, 3- and 5-year OS (m); Sankey diagram of the connection between TNM stage, PTCD1 expression, and survival status (N); *p < 0.05; **p < 0.01; ***p < 0.001

3.2 The association of PTCD1 expression and clinical outcomes

The TCGA-KIRC populations subsequently split into two groups based upon this median level of PTCD1 as the cutoff. Table 1 showed some characteristics of ccRCC populations, and we discovered that elevated PTCD1 expression related to much worse stage (p = 0.0016), grade (p = 0.0020), T stage (p = 0.0080), N stage (p = 0.0247), M stage (p = 0.0018), and survival condition (p < 0.0001). The Kaplan–Meier curves demonstrated that low-PTCD1 group exhibited significantly greater overall survival (OS) (Fig. 2j; p < 0.001), disease specific survival (DSS) (Fig. 2k; p < 0.001), and progress free interval (PFI) (Fig. 2l; p = 0.001) than high-PTCD1 group. The receiver operating characteristic (ROC) curve indicated that PTCD1 had a better reliability in predicting survival prognosis of ccRCC cases (Fig. 2m). The Sankey graph revealed a correlation among pTNM stage, PTCD1 expression, and surviving status, revealing that almost all individuals with low-PTCD1 transcription remained live (Fig. 2n). The subgroups assessment of OS has also been done on age (Fig. 3a, b), gender (Fig. 3c, d), grade (Fig. 3e, f), stage (Fig. 3g, h), T stage (Fig. 3i, j), and M stage (Fig. 3k, l) (all p < 0.05). The findings revealed that overexpression of PTCD1 seemed to have a stable influence on ccRCC individuals with varying clinical features.

Table 1 The relationship between the expression of PTCD1 and various clinicopathological variables in the TCGA databaseFig. 3

The stratified analysis of OS based on clinicopathological characteristics including age (a, b), gender (c, d), grade (e, f), stage (g, h), T stage (i, j) and M stage (k, l), which showed PTCD1 expression had a stable prognosis on ccRCC, respectively (all p < 0.05)

3.3 Nomogram construction and evaluation

The univariate and multivariable Cox regression confirmed that PTCD1 expression and age were potential independent risk factors for ccRCC cases (Fig. 4a, b). PTCD1 transcription and age at diagnosis were used to develop a nomogram in R using the “rms” package to estimate for 1-, 3-, and 5-year OS of ccRCC cases (Fig. 4c). The ROC curve confirmed that PTCD1 expression was more accurate than some other clinical factors (Fig. 4d). The C-index = 0.654 and 1-, 3-, and 5-year curves demonstrated the reliability and accuracy of predictive nomogram (Fig. 4e). Using decision curve analysis revealed that nomogram was more accurate than other clinical factors in evaluating the prognosis of ccRCC (Fig. 4f).

Fig. 4

a Univariate Cox regression analysis; b Multivariate Cox regression analysis; c Nomogram for predicting probability of patients with 1-, 3-, and 5-year OS; d The ROC of PTCD1 expression and clinicopathological characteristics; e Actual and predicted survivals by the calibration curves; f Decision curve analysis of nomogram and other clinical indicators

3.4 Investigation of functionality enriched between low- and high-PTCD1

The most common GO keywords were “neutrophil activation involved in immune response”, “RNA splicing”, “proteasomal protein catabolic process”, “cell-substrate junction” and “focal adhesion” (Fig. 5a). Additionally, KEGG pathways analysis indicated that DEGs were particularly linked to PI3K-Akt signaling, focal adhesion, PD-L1 expression, and PD-1 checkpoint pathway in cancer (Fig. 5b). The circle graph also displayed the top 15 significant GO and KEGG terms, as well as the corresponding genetic enrichment (Fig. 5c, d).

Fig. 5

Functional enrichment analysis between low-PTCD1 group and high-PTCD1 group in TCGA dataset; a Enriched biological process, cellular component, and molecular function; b Enriched Kyoto Encyclopedia of Genes and Genomes pathways; c, d The circle plot of the top 15 most relevant terms in GO analysis and KEGG analysis

3.5 WGCNA and selection of module

To screen the modules linked with PTCD1, co-expression analysis was implemented to generate gene co-expression networks using DEGs datasets. To cluster tumor samples from the DEGs datasets, Pearson’s correlation and average linkage techniques were applied (Fig. 6a). There weren't any suspicious specimens found or discarded. Optimal = 12 (scale-free R2 = 0.9) was selected to guarantee that scale-free nets were constructed inside the DEGs dataset. With a threshold of 0.25 and a minimum module number of 30, seven modules (Fig. 6b) were retained for subsequent analysis. The scatterplot would be used to evaluate potential correlations among modules and traits, and it revealed that the pink component (cor = 0.62, p = 3.2e−67) was strongly linked with expression levels of PTCD1 (Fig. 6c).

Fig. 6

Identification of the module related with PTCD1 in DEGs dataset. a Clustering dendrograms of samples as well as traits; b Cluster dendrogram of co-expression network modules based on the 1-TOM matrix; c High-PTCD1 expression was closely related to the pink module; d Co-expression network of PTCD1 in the pink module; e Dot heatmap of gene correlation in co-expression network

3.6 Generation of a co-expression connection and correlation

Researchers selected DEGs in the pink module that were closely related with PTCD1 to build a co-expression network using Cytoscape 3.6.0 software (Fig. 6d). The dot heatmap displayed its power of each gene's communication (Fig. 6e), revealing that PTCD1 had a favorable co-expression connection with MT-ND2 (r = 0.335), MT-CO1 (r = 0.316), MT-CO2 (r = 0.404), MT-CYB (r = 0.366), MT-ATP6 (r = 0.353), and MT-ND4 (r = 0.378), and a negative co-expression partnership with IARS (r = −0.316) and LRPPRC (r = −0.443) (all p < 0.001; Fig. 7a–h).

Fig. 7

The association of PTCD1 with top eight core genes including IARS2 (a), LRPPRC (b), MT-ND2 (c), MT-CO1 (d), MT-CO2 (e), MT-CYB (f), MT-ATP6 (g) and MT-ND4 (h)

3.7 Potential association of PTCD1 transcription with immune infiltration

We collected the results of 530 ccRCC patients computed using various algorithms and analyzed all immune cell subtypes, which revealed that the percentage of partial infiltrating cell subtypes differed significantly between the two groups (Fig. 8). The ssGSEA analysis found that NK CD56 bright cell and Treg had a higher enrichment score in the high-PTCD1 group, while T helper cells, eosinophils, Th17 cells, macrophages, Th2 cells and neutrophils had a lower proportion in the high-PTCD1 group (all p < 0.05; Fig. 9a–h).

Fig. 8

Using different algorithms to calculate immune cell infiltration in 530 ccRCC cases, demonstrating that the proportion of immune cell infiltration was significantly different between the two groups

Fig. 9

The enrichment scores of eosinophils (a), macrophages (b), neutrophils (c), NK CD56 bright cell (d), T helper cells (e), Th17 cells (f), Th2 cells (g), Treg (h) between high-PTCD1 group and low-PTCD1 group; The immune checkpoint (i), EMT-related genes (j) of the high- and low-PTCD1 group for ccRCC patients in the TCGA cohorts. *p < 0.05; **p < 0.01; ***p < 0.001

3.8 The relationship of PTCD1 expression, immune checkpoints and EMT-related genes

Researchers explored the level of immune checkpoints-related genes transcription and identified that some markers (CTLA4, LAG3, CD27, TNFRSF18, CD244, TMIGD2, LAIR1, TIGIT, TNFRSF14, TNFRSF9, PDCD1, IDO2, TNFSF9, LGALS9, CD70, ADORA2A, BTNL2, TNFRSF25, TNFSF14, and TNFRSF8) were elevated in the high-PTCD1 group, and some biomarkers (NRP1, KIR3DL1, PDCD1LG2, IDO1, TNFSF18, and CD274) were negatively regulated in the high-PTCD1 group, indicating the existence of immunological and fatigued phenotype (Fig. 9i). The expressions of EMT-related genes in two groups were also analyzed and identified ENO2, LGALS1, VEGFA in the low-PTCD1 group was upregulated, and SFRP1, LOX, BDNF were downregulated, indicating some genetic epigenetic changes in the high-PTCD1 group (Fig. 9j, all p < 0.05). Based on these results, we observed that individuals in 2 groups exhibited dramatically distinct patterns of immune infiltration and EMT characteristics, which could result in varied prognosis.

3.9 Prediction of anti-tumor drug sensitivity and immunotherapy response

The results indicated that the high-PTCD1 group was more sensitive to GSK690693, Sorafenib, Cisplatin and Axitinib, as shown in Fig. 9a–d. Moreover, the low-PTCD1 group was more sensitive to Saracatinib, Pazopanib, Sunitinib and Rapamycin, as shown in Fig. 10e–h. The TCIA dataset was utilized to construct the IPS of ccRCC cases, which was a strong predictor of responsiveness to anti-CTLA-4 and anti-PD-1. The results identified that individuals in the high-PTCD1 group had a substantially lower reaction than patients in the low-PTCD1 group for CTLA-4 positive or both negative, which strongly indicated that patients with high-PTCD1 level would have a poorer immunological reaction (Fig. 10i, j). These findings will help us choose specific drugs based on anti-tumor drug sensitivity.

Fig. 10

Analysis of anti-tumor drug sensitivity (a–h) and immunotherapy response (i, j) between high- and low-PTCD1 groups; (k) Normal tissues and Tumor tissues (l) on the protein levels of PTCD1 from HPA database

3.10 Immunohistochemistry

The Human Protein Atlas website (https://www.proteinatlas.org) provided IHC staining data, which was a collection focused on proteome, transcriptome, and systems biology that could map tumors, cells, and tissues. HPA achieved IHC staining of PTCD1 in cancer tissues and normal tissues and discovered that its protein levels of PTCD1 of ccRCC cases were considerably higher than normal tissues (Fig. 10k, l).

3.11 Upregulation of PTCD1 promoted the proliferation, migration and invasion of ccRCC cells

We transfected the PTCD1 vector into 786o and ACHN cells, which showed a relatively low expression of PTCD1, to upregulate PTCD1 expression (Fig. 11a, b). We found that upregulation of PTCD1 markedly increased the proliferation and viability of 786o and ACHN cells (Fig. 11c–f). Transwell invasion assay revealed that upregulation of PTCD1 also promoted the invasive ability of 786o and ACHN cells (Fig. 11g, h). Wound-healing assays showed that PTCD1 overexpression significantly increased 786o and ACHN cell migrations, compared to the negative control groups (Fig. 11i, j). Colony formation assays showed that upregulation of PTCD1 distinctly generated more colonies compared with the negative control cells (Fig. 11k, l). Thus, PTCD1 accelerated the proliferation, migration and invasion of ccRCC cells.

Fig. 11

Overexpression of PTCD1 promoted the proliferation, viability, migration and invasion of ccRCC cells in vitro. Expression levels of PTCD1 mRNA in HK-2 and four ccRCC cell lines were detected by qRT-PCR (a). qRT-PCR revealed that PTCD1 was markedly upregulated by pLV-PTCD1 in the 786o and ACHN cells (b). Upregulation of PTCD1 markedly promoted the proliferation (c, d) and viability (e, f) of 786o and ACHN cells; Transwell invasion assays accessing the invasive ability of 786o and ACHN cells (g, h); Wound-healing assays accessing the invasive ability of 786o and ACHN cells (i, j); Colony formation assays showed the more colonies in pLV-PTCD1 cells than LV-NC cells (k, l); Upregulation of PTCD1 increased the level of β‐catenin, and contributed the phosphorylation of AKT at Ser473, GSK-3β at Ser9 (M). *p < 0.05; **p < 0.01; ***p < 0.001

3.12 PTCD1 was a positive regulator of the Akt/GSK‐3β/β-catenin pathway

To gain insight into the PTCD1-mediated effects in ccRCC, based on the TCGA cohort-associated pathway of enrichment, we assessed the role of PTCD1 in regulating Akt/GSK‐3β/β-catenin pathway in 786o and ACHN cell lines (Fig. 11m). The results identified that upregulation of PTCD1 strikingly increased the level of β‐catenin, which was reported that it can regulate TCF/LEF-mediated transcriptional activity. Upregulation of PTCD1 also increased the phosphorylation of AKT at Ser473 and GSK-3β at Ser9, indicating that PTCD1 promoted the activation of AKT/GSK-3β, which has been proved that it contributed significantly to tumor initiation and progression in ccRCC. Collectively, our findings implied that PTCD1 promotes the Wnt/β-catenin pathway by regulating AKT/GSK-3β.