PRMT6 expression in different human tissues and cancers
To elucidate the differences between PRMT6 in healthy tissues and tumor tissues, we performed a visual analysis of the PRMT6 expression in healthy tissues. As observed in Figure 1a, PRMT6 is expressed at low levels in healthy brain tissues. Then we also analyzed the PRMT6 expression in different tumors and normal tissues. Tumors with significant positive associations included: Adrenocortical carcinoma (ACC), Bladder Urothelial Carcinoma (BLCA), Breast invasive carcinoma (BRCA), Cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), Colon adenocarcinoma (COAD), Esophageal carcinoma(ESCA), Glioblastoma multiforme (GBM), LGG, Liver hepatocellular carcinoma (LIHC), Lung adenocarcinoma(LUAD), Lung squamous cell carcinoma (LUSC), Ovarian serous cystadenocarcinoma (OV), Pancreatic adenocarcinoma (PAAD), Prostate adenocarcinoma (PRAD), Skin Cutaneous Melanoma (SKCM), Stomach adenocarcinoma (STAD), Testicular Germ Cell Tumors (TGCT), Thyroid carcinoma (THCA), Uterine corpus endometrial carcinoma (UCEC), Uterine Carcinosarcoma (UCS). There were also some negatively correlated tumors: Kidney renal clear cell carcinoma (KIRC), Acute Myeloid Leukemia (LAML), Kidney Chromophobe(KICH), Kidney renal papillary cell carcinoma (KIRP).
Analysis of survival prognosis
We used clinical information from the TCGA database to investigate the correlation between PRMT6 and survival prognosis. Forest maps of OS, DSS, and PFI showed that the PRMT6 expression is a risk factor in BLCA, LGG and UCEC, as well as a protection factor in BRCA. In addition, PRMT6 expression was divided into high and low groups, with survival analysis curves. In OS analysis, the high and low PRMT6 expression was significant in LGG and UCEC. In DSS analysis, the high and low PRMT6 expression was significant in LGG, LUAD and UCEC. In PFI analysis, the high and low PRMT6 expression was significant in LGG, COAD and LUAD. These evidence suggest that PRMT6 has important reference for prognosis in tumors.
Multidimensional immune correlation analysis
In order to investigate the effect of PRMT6 on immune level, we analyzed its interaction with different immune infiltrating cells, immune microenvironment, immune subtype and immune checkpoint genes from different perspectives. As we can see from Additional File 1 Figure S1, PRMT6 expression in BRCA, CESC, Head and Neck squamous cell carcinoma(HNSC), KIRC, LGG, LIHC, PAAD, SKCM, Thymoma (THYM), UCEC are significantly correlated with the immune cells existed, indicating that PRMT6 expression is strongly correlated with immune cells in tumors.
Then, we analyzed the relationship between PRMT6 expression and stromal cell and immune cell score in tumors. The stronger the correlation, and the higher the score, the more significant the proportion of stromal cells and immune cells in the tumor tissue. As we can see from the Additional file 2 Figure S2, gene expression was negatively correlated with immune cells or stromal cells of score in BLCA, CESC, GBM, LAML, LUSC, OV, Pheochromocytoma and Paraganglioma (PCPG), PRAD, Sarcomav(SARC), STAD, TGCT, THCA, THYM and UCEC. Interestingly, it's only positive in LGG, which can further prove the intrinsic relationship between PRMT6 expression in g and immune cells.
Also, we analyzed the expression relationship between PRMT6 and 47 common immune checkpoint genes. From Figure 3b, we can see that the PRMT6 expression in various tumours is strongly correlated with immune checkpoint genes.
Finally, we used the TISIDB online tool to analyze the relationship between gene expression and immune or molecular subtypes (Additional file 3 Figure S3). The results show that the expression of LGG is significantly higher than others’.
Mutation correlation analysis
We calculated TMB in each cancer tumor, and we can see that from the radar map in ACC, THCA, PRAD, PAAD, OV, LGG, COAD and CESC have significantly correlation. In the MSI radar map, SKCM, PAAD, LUAD, KIRC, Lymphoid Neoplasm Diffuse Large B-cell Lymphoma (DLBC) and COAD have significant differences. In the MSI, COAD, DLBC KIRC, LUAD, PAAD, SKCM have statistically significant differences.
Various clinical indicators correlation analysis
To further elucidate the potential clinical value of PRMT6 in glioma, we analyzed various clinically common indicators including age, tumor grade, IDH mutation status, 1p19q co-deletion status, chemotherapy status. As can be seen from Figure 5, the PRMT6 expression level in TCGA and CGGA databases was significantly correlated with various clinically common risk factors.
Independent prognostic risk factor analysis
We considered high and low PRMT6 expression as an independent risk factor. First, survival analysis curves were performed based on TCGA and CGGA data, showing that the high and low PRMT6 expression significantly affected the prognosis of patients. Then, we constructed the nomogram by screening for various independent risk factors. We used TCGA database as the training set and CGGA database as the validation set for external validation. As from the results of the nomogram, PMRT6 is potentially valuable as an independent prognostic risk factor. In order to test the accuracy of the nomogram model, we calculated the C-index, which was 0.84 (95%CI:0.865-0.815) in the nomogram of TCGA, and 0.772 (95%CI:0.801-0.743) in the nomogram of CGGA. The calibration curves of the 1-year and 3-year survival rates of the two models were drawn respectively, and the results showed that the two models had better validation performance.
Gene enrichment results
According to the results of KEGG and GO enrichment analysis, the role of PRMT6 in gliomas is mainly related to the regulation of cell cycle, the involvement of DNA damage and repair, and the conduction of some signaling pathways. Table 2 reveals some representative pathways and related functions of enrichment.
Table 2: Gene enrichment results
Gene set names
|
NOM p-val
|
FDR q-val
|
KEGG gene set
|
|
|
KEGG_BLADDER_CANCER
|
0
|
0.038
|
KEGG_SMALL_CELL_LUNG_CANCER
|
0.002
|
0.038
|
KEGG_PANCREATIC_CANCER
|
0.006
|
0.044
|
KEGG_SYSTEMIC_LUPUS_ERYTHEMATOSUS
|
0.002
|
0.046
|
KEGG_P53_SIGNALING_PATHWAY
|
0
|
0.031
|
KEGG_MISMATCH_REPAIR
|
0
|
0.045
|
KEGG_NUCLEOTIDE_EXCISION_REPAIR
|
0.006
|
0.043
|
KEGG_ECM_RECEPTOR_INTERACTION
|
0.008
|
0.047
|
KEGG_PYRIMIDINE_METABOLISM
|
0.002
|
0.034
|
KEGG_GLUTATHIONE_METABOLISM
|
0.002
|
0.049
|
KEGG_AMINO_SUGAR_AND_NUCLEOTIDE_SUGAR_METABOLISM
|
0
|
0.045
|
GO gene set
Cell cycle regulation
|
|
|
GO_CELL_CYCLE_G1_S_PHASE_TRANSITION
GO_POSITIVE_REGULATION_OF_CELL_CYCLE_PHASE_TRANSITION
|
0
0
|
0.029
0.033
|
GO_REGULATION_OF_CELL_CYCLE_PHASE_TRANSITION
|
0.002
|
0.032
|
GO_REGULATION_OF_DNA_TEMPLATED_TRANSCRIPTION_IN_RESPONSE_TO_STRESS
|
0
|
0.032
|
GO_REGULATION_OF_POSTTRANSCRIPTIONAL_GENE_SILENCING
|
0
|
0.037
|
GO_REGULATION_OF_TRANSCRIPTION_FROM_RNA_POLYMERASE_II_PROMOTER_IN_RESPONSE_TO_HYPOXIA
|
0.002
|
0.038
|
GO_SIGNAL_TRANSDUCTION_INVOLVED_IN_CELL_CYCLE_CHECKPOINT
|
0
|
0.041
|
GO_MITOTIC_CELL_CYCLE_CHECKPOINT
|
0.002
|
0.035
|
DNA damage and repair
|
|
|
GO_DNA_DAMAGE_RESPONSE_DETECTION_OF_DNA_DAMAGE
|
0
|
0.042
|
GO_DNA_DAMAGE_RESPONSE_SIGNAL_TRANSDUC GO_DNA_SYNTHESIS_INVOLVED_IN_DNA_REPAIR
|
0
0.002
|
0.042
0.041
|
GO_G1_DNA_DAMAGE_CHECKPOINT
|
0.004
|
0.047
|
GO_NUCLEOTIDE_EXCISION_REPAIR_DNA_GAP_FILLING
|
0.004
|
0.045
|
GO_SIGNAL_TRANSDUCTION_IN_RESPONSE_TO_DNA_DAMAGE
Signal transduction
GO_REGULATION_OF_SIGNAL_TRANSDUCTION_BY_P53_CLASS_MEDIATOR
GO_SIGNAL_TRANSDUCTION_BY_P53_CLASS_MEDIATOR
|
0
0.004
0.002
|
0.04
0.042
0.039
|
GO_TUMOR_NECROSIS_FACTOR_MEDIATED_SIGNALING_PATHWAY
|
0
|
0.042
|
Gene sets with NOM p-val and FDR q-value<0.05 are considered as significant.
Immunohistochemical results
Immunohistochemical results showed that PRMT6 protein was positively expressed in the nucleus and was brownish yellow or brown in color. The positive rate was 87.5% in glioma and 25% in normal tissue, with statistical difference, and a P value of 0.0055. Additionally, the PRMT6 expression in normal tissues, LGG and GBM was significantly different, with a significant upward trend. Based on the obtained clinical information, we drew a survival analysis curve, which further verified the significant correlation between high and low PRMT6 expression and prognosis.
Table 3: Expression of PRMT6 in glioma and normal tissues
Tissue types
|
Number of cases
|
PRMT6
|
P value
|
Negative
Expression
|
Low
Expression
|
High
Expression
|
Glioma tissue
|
32
|
4
|
16
|
11
|
0.0055
|
Normal tissue
|
4
|
3
|
1
|
0
|