3.1 Construction of a Cuproptosis-related lncRNA signature
A total of 258 patients with OC were randomly assigned to the training (n = 132) and testing groups (n = 126) in a 1:1 ratio. In the two cohorts, 68.62% of the patients aged < 65 years old and 93.09% of the population had advanced OC (stage III−IV) and 85.64% of the patients were grade 3, indicating that the tumor was highly malignant; however, there were no significant differences in clinical characteristics (age, grade, and stage) between the two cohorts (Table 1, P > 0.05).
Table 1
Comparison of clinical characteristics between two cohorts of ovarian cancer
Variates
|
Type
|
Total
N (%)
|
Testing group
N (%)
|
Training group
N (%)
|
P value
|
Age
|
<=65
|
258 (68.62)
|
126 (67.02)
|
132 (70.21)
|
0.5784
|
|
> 65
|
118 (31.38)
|
62 (32.98)
|
56 (29.79)
|
|
Grade
|
1
|
1 (0.27)
|
0 (0.00)
|
1 (0.53)
|
0.0875
|
|
2
|
42 (11.17)
|
27 (14.36)
|
15 (7.98)
|
|
|
3
|
322 (85.64)
|
155 (82.45)
|
167 (88.83)
|
|
|
unknow
|
11 (2.93)
|
6 (3.19)
|
5 (2.66)
|
|
Stage
|
I
|
1 (0.27)
|
0 (0.00)
|
1 (0.53)
|
0.5079
|
|
II
|
22 (5.85)
|
13 (6.91)
|
9 (4.79)
|
|
|
III
|
292 (77.66)
|
146 (77.66)
|
146 (77.66)
|
|
|
IV
|
58 (15.43)
|
26 (13.83)
|
32 (17.02)
|
|
|
unknow
|
3 (0.80)
|
3 (1.60)
|
0 (0.00)
|
|
Information about Cuproptosis-related genes and 16876 lncRNAs of OC was obtained from the TCGA database. Furthermore, lncRNAs associated with Cuproptosis-related genes were screened using co-expression analysis (Fig. 1A). In total, 430 Cuproptosis-related lncRNAs were identified (Table S1).
To investigate the relationship between Cuproptosis-related lncRNAs and survival outcomes in patients with OC, lncRNAs with prognostic significance were further screened. First, a univariate Cox regression analysis was performed, and 16 Cuproptosis-related lncRNAs with prognostic significance were selected (Table S2, Figure S1). Fourteen Cuproptosis-related lncRNAs were screened using LASSO regression and cross-validation (Fig. 1B, 1C). Next, eight Cuproptosis-related lncRNAs that were closely related to the prognosis of OC were confirmed using multivariate Cox regression analysis (Table S3). Among them, five lncRNAs (AL732292.2, LINC00996, AC025287.2, AC022893.3, and SUCLG2-AS1) were protective indicators in patients with OC, while the other three lncRNAs (AC245041.1, AL391832.3, and AC019080.5) were risk factors (Table S2, Fig. 1D).
3.2 The risk score can serve as an independent prognostic factor and guide the prediction of clinical outcome for TCGA-OV samples
The patients were divided into high- and low-risk groups in the training and testing groups according to the cut-off criteria. Figure 2 shows the expression, risk score, and survival status (dead/alive) of Cuproptosis-related lncRNAs from top to bottom in each sample (Fig. 2A-C, 2E-G). The results showed that the expression of AC245041.1, AL391832.3, and AC019080.5 was higher than in the high-risk group in both the training and testing groups. The higher the risk score for Cuproptosis-related lncRNAs, the worse the prognosis for patients with OC. KM analysis showed that OS of patients with OC in the high-risk group was worse in both cohorts (P < 0.05, Fig. 2D, 2H).
To further construct the prognosis model for patients with OC, univariate and multivariate analyses were used to study the predictive ability of Cuproptosis-related lncRNAs. The results showed that Cuproptosis-related lncRNAs were independent risk factors for the prognosis of patients with OC (Fig. 3A, 3B). The area under the curve (AUC) of the risk score was 0.692 at 1 year, 0.648 at 3 years, and 0.651 at 5 years, indicating that Cuproptosis-related lncRNA has a good ability to predict the prognosis of patients with OC (Fig. 3C).
We compared the predictive ability of clinical parameters (grade, stage, and age) and risk scores for OS in patients with OC. The AUC of the risk score and age were similar (0.692 vs. 0.710), which was better than that of the grade and stage in predicting OS in patients with OC (Fig. 3D). The C-index showed that the risk score had the highest predictive accuracy (Fig. 3E). These results suggested that the risk score of Cuproptosis-related lncRNAs can be used as a reliable predictor of OS in patients with OC.
The clinical parameters (grade, stage, and age) and risk score were used to draw a nomogram to predict the survival outcome of patients with OC. The total score on the nomogram was the sum of the corresponding scores assigned to each risk factor. The higher the score, the worse the prognosis of the patients with OC (Fig. 3G). Calibration curves confirmed that the prognostic model had a good ability to predict the 1-, 3-, and 5-year OS (Fig. 3F).
3.3 PCA verified the prediction ability and the grouping power of the prognostic signature
PCA was used to verify whether Cuproptosis-related lncRNAs of prognostic significance could distinguish between the high- and low-risk groups of patients with OC. The results showed that PCA of Cuproptosis-related lncRNAs (Fig. 4B), Cuproptosis-related genes (Fig. 4C), and all genes (Fig. 4D) could not distinguish between the high- and low-risk groups. LncRNAs, which were used to construct prognostic models, were able to better distinguish between the high- and low-risk groups (Fig. 4A).
3.4 Functional enrichment analysis and immune function analysis
By comparing gene expression between the high- and low-risk groups, we further screened 268 DEGs for GO analysis and KEGG analysis (Table S4). GO analysis showed that the functions of these DEGs were mainly related to skeletal system development, collagen-containing extracellular matrix, receptor-ligand activity, and signaling receptor activator activity (Fig. 5A). KEGG analysis showed that DEGs were mainly enriched in the PI3K-Akt signaling pathway and neuroactive ligand-receptor interactions (Fig. 5B).
To explore whether Cuproptosis was related to the immune function of patients with OC, we studied differences in immune function between the high- and low-risk groups. The results suggested significant differences in Type_I_IFN_Reponse, Type_II_IFN_Reponse, T_cell_co-inhibition, T_cell_co-stimulation, HLA, APC_co_stimulation, cytokines and cytokine receptors (CCR), and inflammation between the two groups, and all were more enriched in the high-risk group than in the low-risk group. Furthermore, the difference in CCR and Type_I_IFN_Response was the most significant between the two groups (P < 0.001, Fig. 5C). These results suggest a significant correlation between Cuproptosis and immune function in OC.
3.5 Significance of the risk model in drug therapy
A high TIDE score indicates that the curative effect of immune checkpoint blockade (ICB) treatment was poor, and the survival time was short after ICB treatment [4]. The TIDE score was higher in the high-risk group, suggesting that this group of patients may have had a worse response to ICB treatment (P > 0.05, Fig. 6A).
Therefore, we further investigated the sensitivity of patients with OC in the high- and low-risk groups to 19 antitumor drugs (Figs. 6 and 7). Patients in the high-risk group were more sensitive to THZ-2-49, (5Z)-7-Oxozeaenol, WH-4-023, A-770041, saracatinib, pazopanib, dasatinib, cytarabine, CGP-60474, BEZ235, and AP-24534 (P < 0.05, Fig. 6B-L). Patients in the low-risk group were more sensitive to TAK-715, rTRAIL, CP724714, NSC-207895, PD-0325901, MP470, gefitinib, and FH535 (P < 0.05, Fig. 7A-7H). These results indicated that the risk model could distinguish the effects of antitumor therapy in patients with OC.
3.6 TMB of the Cuproptosis-related lncRNAs prognostic marker in patients with OC
TMB is related to the sensitivity of ICB treatment; therefore, we further explored TMB in the high- and low-risk groups. The results showed that the TMB in the low-risk group was higher than in the high-risk group (P < 0.05, Fig. 8E), indicating that patients in the high-risk group were less sensitive to ICB treatment. The waterfall map showed that 89.16% of the samples in the low-risk group contained mutations, and the three genes with the highest mutation frequencies were TTN (57%), TP53 (39%), and MUC16 (36%) (Fig. 8A); 90.26% of the samples in the high-risk group had mutations, and the three genes with the highest mutation frequencies were TTN (45%), TP53 (45%), and MUC16 (25%) (Fig. 8B). In particular, the frequency of the common cancer mutation gene, MUC16, was higher in the low-risk group, while the mutation frequency of the tumor suppressor gene, TP53, was higher in the high-risk group. KM analysis showed that the OS of patients with low TMB was shorter, and the prognosis was worse than that of patients with low TMB (P < 0.05, Fig. 8C), which was consistent with the conclusions of previous studies. In addition, the prognosis of the high-risk group was worse in the high TMB and low TMB groups (P < 0.05, Fig. 8D).