ORC6, a novel prognostic biomarker, correlates with T regulatory cell infiltration in prostate adenocarcinoma: a pan-cancer analysis

We obtained the ORC6 mRNA expression levels and clinical data of TCGA and GTEx cohorts from the UCSC Xena database (https://​xenabrowser.​net/​datapages/​). p-values < 0.05 (two-tailed) were regarded as statistically significant (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001). Then, we downloaded the ORC6 mRNA expression data for diverse cancer cell lines from the Cancer Cell Line Encyclopedia (CCLE) database (https://​portals.​broadinstitute.​org/​ccle/​data) and the DNA copy number and methylation information from the cBioPortal database (https://​www.​cbioportal.​org/​).

We further evaluated the ORC6 protein levels based on the IHC staining data provide by the Human Protein Atlas (HPA) database (https://​www.​proteinatlas.​org/​). Subcellular localization of ORC6 was observed in the GeneCards database (https://​www.​genecards.​org/​). The data from the STRING database (https://​string-db.​org/​) was built for the protein–protein interaction (PPI) network.

In order to explore the association between ORC6 expression and prognostic information, Kaplan–Meier analysis of the TCGA datasets was performed. Four survival indicators, including the overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI), and progression-free interval (PFI), were enrolled in the analysis. We set up univariate Cox regression analyses to evaluate the prognostic significance of ORC6 in predicting these four survival indicators in these 33 types of cancers. The results of the regression analyses are shown using a forest plot.

The TMB analysis was conducted with the R package (edgeR) using the human pan-cancer somatic data (MAF data) from TCGA database. The MSI score was used as per a published study [16]. Both TMB and MSI were calculated using the Pearson’s method. The “Gene_Corr” module of TIMER2 was utilized to analyze MMR gene expression levels in the TCGA database. Five important MMR genes, including MutL protein homolog 1 (MLH1), MutS protein homolog 2 (MSH2), MutS homologue 6 (MSH6), epithelial cell adhesion molecular (EPCAM), and PMS1 homolog 2 (PMS2) were selected for the correlation analysis. The correlation degree was calculated with purity-adjusted Spearman and plotted in a heatmap.

First, we used the data from the TCGA database to explore the potential biological and molecular functions of ORC6 via both the Gene Set Enrichment Analysis (GSEA) and Gene Set Variation Analysis (GSVA). The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database was selected for GSEA enrichment analyses with the R package “clusterProfiler” [17‐20]. Then, we screened and demonstrated the top 20 most significant positive correlated pathways. In addition, we performed the GSVA with the R package “GSVA” using hallmark pathways from the MSigDB database (https://​www.​gsea-msigdb.​org/​gsea/​msigdb/​index.​jsp).

The ORC6 expression stratified by the immune subtypes across cancers was investigated in the TISIDB database (http://​cis.​hku.​hk/​TISIDB/​). The influence of ORC6 expression on immune cell infiltration was analyzed using the datasets from the TIMER2 database (http://​timer.​cistrome.​org/​). Cancer-associated fibroblasts, tumor endothelial cells, T regulatory (Treg) cells, and CD8⁺ T cells were selected for detailed analysis. The EPIC, MCP-counter, xCELL, CIBERSORT, CIBERSORT-ABS, quanTIseq, and TIMER algorithms were utilized to evaluate endemic tumor cell types in TCGA. The Spearman correlation analysis between the ORC6 expression and immune checkpoint-associated genes was conducted using the TCGA pan-cancer data and visualized using a heatmap.

A total of 19 formalin-fixed and paraffin-embedded prostate adenocarcinoma tumor tissue samples were rehydrated and incubated with the anti-ORC6 (1:400; Genetex, USA), anti-FOXP3 (1:400; Genetex, USA), and anti-CD4 (1:400; Thermo Fisher Scientific, USA) antibodies in a humid box at 4 °C. Three representative 500 × 430 µm areas with more than 50% tumor cell and more than 30 CD4 + cells were enrolled for staining analysis. The expression level of ORC6 was evaluated based on the tissue immunostaining score (TIS), which was defined as the product of the intensity score (IS) and quantity score (QS) (i.e., TIS = IS × QS). The positive staining of ORC6 in tumor cells and FOXP3 in CD4 + cells was regarded as the staining percentage. The staining percentage scores and the staining intensity scores were calculated as previously published [21]. For the FOXP3 and CD4, positive staining cells were counted for the Treg and CD4 + T cells. Mann–Whitney U test was applied to calculate the relationship between the ORC6 expression and Treg and CD4 + T cell number with the GraphPad Prism 8 (GraphPad Software, USA). All p-values < 0.05 (two-tailed) were regarded as statistically significant, and denoted as ∗ p < 0.05, ∗  ∗ p < 0.01, ∗  ∗  ∗ p < 0.001, and ∗  ∗  ∗  ∗ p < 0.0001, respectively. We obtained written informed consent from all patients and implemented all procedures under the Declaration of Helsinki.

We first investigated the ORC6 status in pan-cancer by analyzing the available data from TCGA, encompassing 33 most common types of cancers, including BLCA, CESE, and DLBC (Table 1). The increased ORC6 expression was observed in 29 types of cancers, including BLCA, CESE, and PRAD, compared to the corresponding normal adjacent tissues while decreased ORC6 expression was observed only in LAML (Fig. 1A). The highest ORC6 expression levels were found in TGCT, CESC, and UCS (Fig. 1B). According to the ORC6 expression levels in normal human tissues based on the GTEx database, ORC6 was mainly expressed in the bone marrow, testis, and spleen (Fig. 1C). According to the information about distinct cell lines extracted from the CCLE database, the highest ORC6 expression levels were found in ALL (acute lymphoblastic leukemia), normal breast (NB), and DLBC cells (Fig. 1D). Further analysis of TCGA data revealed that ORC6 expression was significantly increased in tumor tissues versus adjacent normal tissues in 18 types of cancers, including BLCA, BRCA, and PRAD (Supplementary Fig. 1). Further comparison of the ORC6 expression according to the TCGA database revealed that ORC6 expression was significantly increased in higher-stage tumor tissues than in lower-stage tumor tissues in 11 types of cancers, including ACC, KICH, and LUAD; however, it was decreased in OV and SKCM tissues (Fig. 2).

Genetic alterations in ORC6 were investigated in the cBioPortal database. Patients with PRAD and SARC harboured a high frequency of gene alterations, among which gene amplification was most commonly observed (Supplementary Fig. 2A). Additionally, the ORC6 mRNA expression level was positively correlated with copy number alteration (CNA) in 21 types of cancers, including BRCA, PRAD, and UCS (Supplementary Fig. 2B). Moreover, the DNA methylation level of the ORC6 promoter was negatively correlated with ORC6 mRNA expression level in DLBC, ESCA, PCPG, PRAD, TGCT, THCA, and UCS (Supplementary Fig. 2C).

Data regarding ORC6 protein levels in various tumors and normal tissues were obtained from the HPA database (Supplementary Fig. 3). The ORC6 protein level was highest in head and neck cancer and testis cancer but lowest in renal cancer (Supplementary Fig. 3A). In normal tissues, the ORC6 was overexpressed in the bone marrow, lymph node, stomach, duodenum, tonsil, colon, pancreas, and testis (Supplementary Fig. 3B). Typical IHC staining figures of ORC6 in 17 pairs of tumors (including BLCA, BRCA, and PRAD) and corresponding normal tissues were shown in Supplementary Fig. 4. Tissues of normal bladder, breast, cervix, colon, oral tissue, kidney, cerebral cortex, liver, lung, ovary, pancreas, prostate, rectum, stomach, testis, thyroid, and endometrium had negative or moderate ORC6 IHC staining, while the corresponding tumor tissues had moderate or strong staining. These results were consistent with the results of ORC6 mRNA expression data from the TCGA database. The ORC6 was mainly located in the nucleus (Supplementary Fig. 3C). Additionally, the PPI network analysis using the tool STRING identified that ORC6 closely interacted with ORC1-5, CDT1, CDC6, MCM4, MCM5, and MCM7 (Supplementary Fig. 3D).

Subsequently, we estimated the survival indicators, including OS, DSS, DFI, and PFI. The OS analysis demonstrated that ORC6 expression level was as an unfavorable indicator for patients with 18 types of cancers (e.g., KIRC, KIRP and PRAD), and a protective marker only for patients with OV and THYM (Fig. 3). Higher expression of ORC6 was significantly associated with worse prognosis in DSS for patients with 17 types of cancers (e.g., ACC, LGG and PRAD), while lower expression of ORC6 was only negatively correlated with the prognosis of COAD, OV, and THYM (Supplementary Fig. 5). According to DFI analysis, ORC6 high expression level was as an unfavorable indicator for patients with BRCA, COAD, KIRP, LIHC, LUAD, PAAD, PRAD, SARC, and THCA, and a protective marker for patients with OV (Supplementary Fig. 6). Finally, according to PFI analysis, ORC6 high expression level acted as an unfavorable indicator for patients with 23 types of cancers (e.g., KIRP, LIHC and PRAD), and as a protective marker only in patients with GBM, OV, STAD, and THYM (Supplementary Fig. 7).

To further explore the influence of ORC6 on OS, the univariate Cox regression analysis was performed. The results indicated that high ORC6 expression level was simultaneously associated with low OS and DSS in 13 types of cancers (e.g., KIRC, KIRP and PRAD), while low ORC6 expression was associated with low OS and DSS in patients with OV (Fig. 4A-B). Based on DFI analysis, high ORC6 expression was as an unfavorable indicator for patients with BRCA, KIRP, LIHC, PAAD, PRAD, SARC, and THCA, but a protective marker in OV (Fig. 4C). Finally, the high expression level of ORC6 was associated with a decreased PFI in 13 types of cancers, including KIRP, LIHC, and PRAD (Fig. 4D). In summary, ORC6 significantly influenced all survival metrics of only five types of cancers (i.e., BRCA, KIRP, LIHC, PAAD, and PRAD).

ORC6 expression level was positively correlated with TMB in 10 types of cancers, including LUAD, PRAD and STAD (Fig. 5A). Additionally, it was significantly positively correlated with MSI in 10 types of cancers, including PRAD, SARC and STAD, but negatively correlated with MSI in DLBC (Fig. 5B). Five MMR gene expression levels were significantly positively correlated with ORC6 expression level in most cancers analyzed (e.g., MLH1: 65.6%; MSH2: 96.9%; MSH6: 90.6%; PMS2: 71.9%; EPCAM: 62.5%) (Fig. 5C).

The potential biological pathways associated with ORC6 was predicted through the KEGG pathway analysis, and the top 20 pathways are shown in Fig. 6 and Supplementary Fig. 8. The high ORC6 expression was significantly associated with the cell cycle and DNA replication related pathways (Fig. 6 and Supplementary Fig. 8). It is noteworthy that ORC6 was also involved in the MMR pathway in 14 types of cancers, including ESCA, GBM and PRAD (Fig. 6). These results indicate a potential role of ORC6 in adjusting the tumor microenvironment. The GSVA score revealed that ORC6 was positively correlated with some cell proliferation pathways (e.g., G2M checkpoint, E2F targets and MYC targets v1–2), “DNA Repair” pathway and “unfolded protein response” pathway in almost all cancers (Fig. 7). These pathways have been identified to be correlated with the advanced stage of cancers and may benefit from immunotherapy [22, 23]. In addition, ORC6 was negatively correlated with several immune pathways (e.g., IL2-STAT5 signaling, inflammatory response and IL6-JAK-STAT3 signaling) in the majority of cancers (Fig. 7).

Six immune subtypes (e.g., C1: wound healing; C2: IFN-γ dominant; C3: inflammatory; C4: lymphocyte depleted; C5: immunologically quiet; C6: TGF-β-dominant) presented significantly different ORC6 expression levels in 15 types of cancers, including BRCA, LIHC, LUAD and PRAD (Supplementary Fig. 9). No differences in ORC6 expression levels in immune cells were observed in the other types of cancers.

According to the TCGA database, the ORC6 expression level was inversely related to the infiltration level of cancer-related fibroblasts in seven types of cancers, including BRCA, LUSC, STAD, and TGCT (Fig. 8A). In contrast, a negative correlation was found in KICH (Fig. 8A). Additionally, ORC6 expression level was negatively correlated with tumor endothelial cell infiltration in 11 types of cancers, including ESCA, KIRC, LUAD, LUSC and STAD, while a positive correlation was observed in LGG (Fig. 8B). Furthermore, a negative correlation was detected between ORC6 expression level and Treg cell infiltration in ESCA and LUSC, whereas a positive correlation was observed in PRAD (Fig. 8C). Finally, a positive correlation was found between ORC6 expression level and CD8 + T cell infiltration in KIRC and UVM (Fig. 8D). Furthermore, to validate the correlation between ORC6 expression and Treg cell infiltration, IHC staining of FOXP3, which serves as a lineage specification factor of Treg cells, was performed in tumor samples from 19 patients with prostate adenocarcinoma. Since the median TIS of ORC6 was 6, we stratified the patients into ORC6 TIS ≤ 6 group and TIS > 6 group. The number of Treg in ORC6 TIS > 6 group is more than that of ORC6 TIS ≤ 6 group (mean 7.43 vs 5.47, p = 0.021, Fig. 8E), while there is no difference of CD4 + T cell numbers between the two groups (mean 51.57 vs 50.69, p = 0.757, Fig. 8F). Of note, the FOXP3 + /CD4 + ratio was also higher in ORC6 TIS > 6 group than that of ORC6 TIS ≤ 6 group (mean 0.14 vs 0.11, p = 0.757, Fig. 8G). Representative staining images are shown in Fig. 8H–K.

The association between ORC6 expression and immune related genes expression levels was then evaluated. ORC6 expression level was significantly correlated with the expression level of a majority of immunosuppressive and immunostimulatory markers in BRCA, DLBC, KIRC, LIHC, LUSC, OV, PAAD, PRAD, THCA, and UVM (Fig. 9 and Supplementary Fig. 10). Interestingly, in most tumor types, immunosuppression-related genes, especially TGFBR1 and PD-L1 (CD274), exhibited a specific correlation with the ORC6 expression (Fig. 9). Altogether, these results indicate that ORC6 may promote immunosuppression in a wide array of cancer types.