1. In the OTUD gene family, only YOD1 was stably and highly expressed in PAAD
In order to explore the prognosis and potential clinical value of different OTUD gene families in PAAD patients, we firstly analyzed the expression of OTUD gene families in PAAD which using ONCOMINE. As shown in Fig. 1 and Table.1, the mRNA expression of YOD1 in PAAD was significantly higher than that in normal tissues (P = 4.43E-8). Other members of the OTUD gene family had no significant difference. Then we used R software to analyze TCGA combined with GTEx data, and the results were shown in Table.2. In PAAD, the expressions of OTUD1 (P = 1.06E-52), YOD1 (P = 2.20E-50) and OTUD5 (P = 1.53E-45) in tumor tissues were higher than those in normal tissues. In addition, we also conducted dual validation in tumor tissues and cell lines through MERAV and GEPIA websites. In MERAV, the expressions of YOD1, OTUD4 and OTUD6b in tumor tissues and cell lines were different from those in normal tissues (Fig. S1). In GEPIA, the expressions of OTUD1, YOD1, and OTUD5 were higher in tumor tissues than in normal tissues (P<0.05) (Fig. S2). Compared with different databases, only YOD1 in the OTUD gene family was found to be stably and highly expressed in PAAD.
Table.1 Differences in the transcription level of the OTUD family in pancreatic cancer and normal pancreatic tissues (ONCOMINE: gene rank < 10%)
Gene
|
Pathology
|
Fold Change
|
P-value
|
tTest
|
References
|
OTUD1
|
PDAC
|
1.786
-1.347
|
8.90E-4
0.029
|
3.313
-1.938
|
Ishikawa Pancreas[24]
Ishikawa Pancreas[24]
|
YOD1*
|
PC
|
2.386
|
4.43E-8
|
6.695
|
Pei Pancreas[36]
|
OTUD3
|
PDAC
|
1.438
|
0.041
|
1.836
|
Grutzmann Pancreas[11]
|
OTUD4
|
\
|
\
|
\
|
\
|
\
|
OTUD5
|
PDAC
|
1.415
|
2.29E-4
|
4.783
|
Buchholz Pancreas[8]
|
OTUD6A
|
\
|
\
|
\
|
\
|
\
|
OTUD6B
|
PAAD
|
1.079
|
6.18E-5
|
4.386
|
TCGA Pancreas[32]
|
ALG13
|
PDAC
|
-1.911
|
0.031
|
-1.911
|
Ishikawa Pancreas[24]
|
PDAC, pancreatic ductal adenocarcinoma; PC, pancreatic cancer; PAAD, pancreatic adenocarcinoma |
Table.2 Differences in the transcription level of the OTUD family in pancreatic cancer and normal pancreatic tissues (TCGA>Ex)
Gene
|
ConMean
|
TreatMean
|
Log2FC
|
P-Value
|
FDR
|
OTUD1*
|
2.34
|
4.82
|
2.47
|
1.06E-52
|
6.36E-52
|
YOD1*
|
1.20
|
2.56
|
1.35
|
2.20E-50
|
8.75E-50
|
OTUD3
|
1.95
|
1.61
|
-0.33
|
7.10E-14
|
9.13E-14
|
OTUD4
|
\
|
\
|
\
|
\
|
\
|
OTUD5*
|
9.94
|
13.71
|
3.76
|
1.53E-45
|
4.15E-45
|
OTUD6A
|
\
|
\
|
\
|
\
|
\
|
OTUD6B
|
1.77
|
2.19
|
0.41
|
1.17E-08
|
1.39E-08
|
ALG13
|
7.57
|
2.89
|
-4.68
|
4.72E-53
|
3.16E-52
|
To further verify the expression of YOD1 gene, we downloaded and analyzed five data sets from the GEO database: GSE15471, GSE16515, GSE28735, GSE71729, and GSE62165 included 278 tumor samples and 186 normal tissues. Results showed that the expression level of YOD1 was significantly increased in PAAD tissues compared with normal tissues. The receiver operating characteristic curve (ROC curve) was used to verify the diagnostic effectiveness, and the results showed that the AUC was in the range of 0.61–0.985 (P = 0.001), with good diagnostic effectiveness (Fig. 2).
In addition, we used the cBioPortal website analysis and found that in PAAD, the percentage of changes of genetic markers in the OTUD gene family varied from 5–15% (OTUD1, 5%; YOD1, 11%; OTUD3, 6%; OTUD4, 6%; OTUD5, 9%; OTUD6A, 4%; OTUD6B, 15%; ALG13, 9%), and it is worth noting that the high proportion of YOD1 mRNA was observed (Fig. 3A).
HPA uses proteomic techniques to identify tumor-type specific protein expression patterns that are differentially expressed in specific types of tumors. After the analysis found that only YOD1 mRNA level was stable and high in PAAD, we tried to explore the protein expression pattern of YOD1 in PAAD by HPA. Immunohistochemical results showed that YOD1 was highly expressed in PAAD compared to normal pancreatic tissue at the protein level (Fig. 3B).
In summary, our results indicate that the transcription and protein expression of YOD1 are stably and highly expressed in PAAD patients.
2. The high expression of YOD1 is an independent factor affecting the poor prognosis of PAAD
We used Kaplan-Meier plotter to analyze 177 PAAD samples and found that for overall survival (OS), YOD1 hazard ratio (HR) = 2.14 (1.29–3.56, P = 0.0025) and for relapse-free survival (RFS), YOD1 HR = 3.37 (1.32–8.6, P = 0.0068) (Fig. 4A). Therefore, the high expression of YOD1 is significantly positively correlated with the poor prognosis of PAAD patients. We also obtained the same results in GEPIA database. Compared with the low-expression group, the OS (HR = 2.2, P = 0.0077) and DFS (HR = 2, P = 0.04) of the group with high expression of YOD1 decreased.
In addition, univariate and multivariate Cox regression analyses were performed on PAAD patient data from the TCGA database. The hazard ratio (HR) of the risk score and 95% CI were 1.138 and 1.025–1.264 (P = 0.016), and 1.121 and 1.010–1.245 (P = 0.032), this indicates that the YOD1 risk model can be used as an independent prognostic factor. The higher the expression of YOD1 is, the worse the prognosis of PAAD patients are (Fig. 4B,C).
These results suggest that YOD1 mRNA expression is significantly associated with prognosis in patients with PAAD, and they may be used as a potential biomarker to predict survival in PAAD patients.
3. Correlation analysis of YOD1 tumor immune microenvironment and stemness index
The PAAD microenvironment has two main characteristics: dense connective tissue hyperplasia and extensive immunosuppression. Due to the specific characteristics of the immune microenvironment, tumor tissue is prone to survival and insensitive to chemoradiotherapy, so it results the poor prognosis[23].
To determine whether the expression of YOD1 affects the composition of the PAAD immune microenvironment, we used single-sample gene set enrichment analysis (ssGSEA). The results in Fig. 5A showed that the high expression of YOD1 was closely related to the decreased infiltration of tumor-infiltrating lymphocytes(TIL) (P<0.01), CD8 + T cells (P<0.01), DCs (P<0.01), and NK cells (P<0.01). The decrease of tumor killer cells will facilitate the immune escape of tumor cells. In immune cell function, the high expression of YOD1 is closely related to the reduction of type II interferon response (Fig. 5B). In the immune subtypes, the high expression of YOD1 is closely related to IFN-γ dominant, which is statistically different from other subtypes (Fig. 5C).
To further confirm whether the level of YOD1 expression affects the immune microenvironment of PAAD, we used the ESTIMATE algorithm to calculate the immune cell score and stromal cell score[20]. Immune cells and stromal cells are the two main non-tumor components. The higher the immune score and stromal score are, the higher the degree of non-tumor cell infiltration in tumor tissue is. Immune cell score (R=-0.26, P = 0.00041) and stromal cell score (R=-0.26, P = 0.00066) showed a negative correlation with the expression of YOD1 (Fig. 5D,E). It was speculated that when YOD1 was overexpressed, the tumor purity was higher.
Cancer stem cells (CSCs) share common biological characteristics with adult stem cells, such as lifespan, self-renewal ability, differentiation, and drug resistance, which play a decisive role in cancer progression [24]. Some studies have found that pancreatic cancer stem cells are highly drug-resistant cells that preferentially drive tumorigenesis and development, and these are associated with poor prognosis of PAAD patients [25]. The results showed that the expression of YOD1 was positively correlated with mRNAsi, R = 0.26 (P = 0.00014) (Fig. 5F). It is suggested that the increased expression of YOD1 is related to the increase of stem cell characteristics. Besides, it promotes the occurrence and development of tumor.
4. Enrichment analysis of YOD1
After understanding the prognostic role of YOD1 in PAAD, we will explore the biological function of YOD1 in PAAD. GO and KEGG enrichment results were displayed in the first 30 bits of NES by bubble diagram. GO analysis showed that the high expression of YOD1 was correlated with biological functions such as metastasis, gemcitabine resistance and transcription (Fig. 6A). KEGG pathway analysis showed that the high expression of YOD1 was correlated with other pathways such as base excise and repair, p53 signaling pathway, RNA degradation, cell-cell adhesion and connection (Fig. 6B). GSEA v.4.0 c2.all.v7.4.symbols.gmt [Curated] and c2.cp.kegg. v7.4.symbols.gmt [Curated] gene sets were analyzed, which satisfied the P-value < 0.05 and false discovery rate (FDR) < 0.25. As shown in Fig. S3, the high expression of YOD1 in PAAD is highly correlated with adhesion, metastasis, invasion, Gefitinib chemotherapy resistance, DNA repair, tumor hypoxic microenvironment and p53, Notch, NRG1, MYC, EGFR, Hippo signaling pathway.
Linkedomics analysis found that 6,505 genes were significantly associated with YOD1 in PAAD(P < 0.05,FDR < 0.01). There were 2968 positively correlated genes and 3537 negatively correlated genes. The first 50 positive and negative co-expressed genes are shown in the heat map. Among genes correlated with YOD1, 23 genes were associated with OS (P < 0.05༉, while 33 genes were negatively correlated with OS (P < 0.05) (Fig. S4).
We used the "clusterProfiler" R package to perform PPI, GO and KEGG enrichment analysis on 23 positively related genes, as shown in Fig. 6C-E. The discovery of the enrichment of cell-related connections and the Hippo pathway is consistent with our previous results. It is proved that YOD1 may regulate the occurrence, development and prognosis of pancreatic cancer through these two key aspects.
To identify upstream molecules that regulate YOD1 mRNA levels in PAAD, we evaluated the miRNA-YOD1 network in PAAD. The result is shown in Fig. 7A. Usually miRNAs inhibit the expression of target genes, so we focused on negative correlations and P < 0.0005 of 12 miRNAs (hsa-mir-675;hsa-miR-150; hsa-miR-202; hsa-miR-1468; hsa-miR-139; hsa-miR-140; hsa-miR-218; hsa-miR-9; hsa-miR-342; hsa-miR-375; hsa-miR-500a; hsa-miR-433). Analysis of the above 13 miRNAs showed that, as shown in Fig. 7B, they could similarly focus on cancer, cell adhesion, p53, TGF-β, Hippo and other cancer-related signaling pathways. And among the above 12 miRNAs, we found that hsa-miR-1468; hsa-miR-139; hsa-miR-140; hsa-miR-218; hsa-miR-9; hsa-miR-342; hsa-miR-375; hsa-miR-500a were negatively associated with poor overall survival of PAAD (Fig. S5 and Table.3). Therefore, these miRNAs may affect the prognosis of PAAD by regulating the expression of YOD1.
Table.3 Overall survival of miRNA that target YOD1 in PAAD.
miRNA
|
Low expression
|
High expression
|
Prognose
|
P-value
|
HR
|
hsa-mir-675
|
n = 62
|
n = 116
|
N/A
|
0.053
|
1.54(0.99–2.4)
|
hsa-mir-150
|
n = 83
|
n = 95
|
N/A
|
0.095
|
0.71(0.47–1.06)
|
hsa-mir-202
|
n = 50
|
n = 128
|
N/A
|
0.19
|
0.75(0.49–1.16)
|
hsa-mir-1468
|
n = 118
|
n = 60
|
Favorable
|
0.02
|
0.59(0.37–0.92)
|
hsa-mir-139
|
n = 133
|
n = 45
|
Favorable
|
0.0047
|
0.46(0.26–0.8)
|
hsa-mir-140
|
n = 96
|
n = 82
|
Favorable
|
0.017
|
0.6(0.39–0.92)
|
hsa-mir-218
|
n = 50
|
n = 128
|
Favorable
|
0.0038
|
0.53(0.35–0.82)
|
hsa-mir-9
|
n = 119
|
n = 59
|
Favorable
|
0.038
|
0.61(0.39–0.98)
|
hsa-mir-342
|
n = 119
|
n = 59
|
Favorable
|
0.0013
|
0.45(0.27–0.74)
|
hsa-mir-375
|
n = 93
|
n = 85
|
Favorable
|
0.012
|
0.59(0.39–0.9)
|
hsa-mir-500a
|
n = 84
|
n = 94
|
Favorable
|
0.0088
|
0.58(0.39–0.88)
|
hsa-mir-433
|
n = 116
|
n = 62
|
N/A
|
0.069
|
0.66(0.43–1.04)
|
5. Function verification of YOD1
To verify the function of YOD1, we first compared the expression levels of YOD1 in 6 PAAD cell lines and normal pancreatic cells. It was found that the expression of YOD1 in 6 PAAD cells was higher than normal pancreatic cells HPDE6-C7 (Fig. 8A). Further, we compared the expression of YOD1 protein in the cancerous tissues and the adjacent tissues 5cm or more away from the cancerous tissues of the six pancreatic cancer patients, and we found that the expression of YOD1 in the cancerous tissues was higher than that in the adjacent tissues (Fig. 8B).
We selected MIA-PaCa-2 with low YOD1 expression to express YOD1 highly, and metastatic PAAD with high YOD1 expression and high evil-doing degree AsPC-1 to knock down YOD1 for the experiment. Transient transfection was performed on the cells to verify the expression of EMT-related factors (Fig. 8C,D), and Transwell was performed (Fig. 8E,F). Changes in the protein levels of epithelial-mesenchymal transition markers, including E-cadherin and Snail1, were consistent with the results of Transwell experiment. All the above indicated that after YOD1 knockdown, the migration ability of AsPC-1 cells was decreased, while the high expression of YOD1 increased the migration ability of MIA-PaCa-2 cells. EdU (5-ethynyl-2 '-deoxyuridine) was used to detect cell proliferation. The results showed that YOD1 was decreased, and the activity and proliferation of AsPC-1 cells were decreased, while YOD1 was increased, and the activity and proliferation of MIA-PaCa-2 cells were increased(Fig. 8G,H).
In conclusion, YOD1 may act as a pro-oncogene and have certain effects on the proliferation and metastasis of PAAD cells.