Among the most common tumors, BC is the tenth most common cancer, and the male-to-female ratio is approximately 3.5 : 1. However, due to the limitations of pathological staging in predicting the prognosis of different patients, it is particularly important to discover a new prognostic feature. Recently, with the advancement of sequencing technology, more and more researchers have begun to explore the prognostic value based on genes. Apoptosis has historically been believed to be the only form of programmed cell death (PCD), and necrosis, which was believed to be an “accidental” type of death not regulated by molecular events, was assumed to be the diametrically opposite modality of cell death compared to apoptosis until necroptosis was discovered as a novel programmed form of necrotic cell death that bears a mechanistic resemblance to apoptosis and a morphological resemblance to necrosis. Due to the important role of necroptosis-related lncRNA in the breast cancer, colon cancer and Gastric Cancer[24, 28, 29], we constructed its signature in BC for the first time to predict the prognosis of BC patients.
We constructed a stable prediction signature with 13 genes, including "RP11-111K18.2", "RP11-262A16.1", "RP11-446N19.1", "NR2F2-AS1", "LINC-PINT", "CTD-2027I19.3", "RP11-291B21.2", "USP30-AS1", "RP5-1184F4.5", "AC006160". 5", "DLEU1-AS1", "SLC25A25-AS1", and "RP11-385D13.3", the areas under the ROC curve of 1 year, 3 years and 5 years were 0.73, 0.79 and 0.80, respectively. According to multiple factors ROC curve, the areas of risk signature was higher than that of age (0.673), pathological stage (0.642), pathological T stage (0.615), pathological N stage (0.621) and Gender (0.452) respectively. Next, we made a prediction model with the risk characteristics, T stage, N stage, and age, and displayed it with a nomogram. The correction curve showed that the prediction model could accurately predict the true survival rate. We further validated the robustness of the risk model in clinicopathological subgroups, the judgment of overall survival rate was no significance in the low clinical stage (P = 0.22), low T stage (P = 0.23), and low pathological grade (P = 1) subgroups due to the limitation of sample size, other subgroups could make significant distinctions, showing the accuracy of the model. We randomized 396 patients into two testing data sets and the overall survival rate of high-risk patients was lower than that of low-risk patients in both testing data sets. It illustrated the applicability of our risk signature.
We analyzed each gene in the risk signature, NR2F2-AS1 induced epithelial-mesenchymal transition in non-small cell lung cancer, promoted prostate cancer cell proliferation by regulating CDK4, and modulated the MAPK pathway to regulate colorectal cancer invasion and metastasis[30–32]. LINC-PINT inhibited the proliferation and migration of laryngeal squamous cell carcinoma by silencing ZEB1, and it can also inhibited the invasiveness of thyroid cancer by downregulating miR-767-5p to induce TET2 expression. USP30-AS1 acted as a ferroptosis-related LncRNA signature to predict the prognosis of BC patients. Li Y's study showed that down-regulation of SLC25A25-AS1 promoted the proliferation of colorectal cancer cells and was resistant to chemotherapy. The other lncRNA included in our signature have not been explored, which will be researched in the future studies by our group.
We performed GSEA enrichment analysis on genes in high and low risk groups, and found that cancer-related pathways and some intercellular adhesion pathways were significantly enriched in high risk groups. The HEDGEHOG signaling pathway was considered to play an important role in the occurrence of BC tumor stem cells. The WNT signaling pathway and the MAPK signaling pathway played an important role in the invasion and spread of BC tissues[38, 39], TGF-β signaling pathway had been implicated in the occurrence of various human diseases including malignant tumors. Paxillin enabled breast tumor invasion by maintaining adherens junctions. Xu X's study showed that the tight junction protein ZONAB was upregulated in BC cell lines and promoted BC invasion, indicating that this high-risk group had a poor prognosis. However, metabolic-related pathways were significantly enriched in the low-risk group. Such as peroxisomes, reactive oxygen species could be produced in peroxisomes, elevated ROS production efficiently inhibited chemo-drug resistance and promoted chemoresistant cell death, and α-linolenic acid intake has a protective effect on the development of BC.
We explored the relationship between immune cell infiltration and necroptosis-related LncRNAs and found that the risk level of BC patients may be potentially affected by immune infiltration. In recent years, more and more researchers had paid attention to the immunotherapy. The clinical impact of checkpoint blockade strategies, providing a survival advantage compared with traditional chemotherapies, had grown considerably, and had been tested in various tumors including melanoma, renal cell carcinoma, non-small cell lung cancer, and urothelial carcinoma over the past several decades. Phase III clinical trial illustrated that pembrolizumab was associated with a lower rate of treatment-related adverse events and with significantly longer OS than chemotherapy as a second-line therapy for platinum-refractory advanced urothelial carcinoma. However, the expression of immune checkpoints was not significantly different between high-risk and low-risk groups. Recent work suggests that MSI may be used as a predictor for immune-checkpoint blockade therapy. Several clinical studies have shown better outcomes among patients with MSI-positive tumors as compared to negative groups when they are treated with inhibitors of programmed cell death 1 (PD-1)[46, 47]. In our study, the microsatellite status was more stable in the high-risk group, so the low-risk group were more effective to immunotherapy, consistented with our risk characteristics predictions.
Finally, we used risk signature to predict patient response to chemotherapy drugs. The low-risk group were more sensitive to methotrexate. In contrast, patients in the high-risk group were more sensitive to cisplatin, docetaxel, paclitaxel and thapsigargin. Taken together, our necroptosis-associated lncRNA signature could accurately predict chemotherapy response in patients with BC, which may be helpful for clinical medical decision-making.
The Thirteen-gene prognostic model can effectively predict the prognosis of patients with BC and may provide a clinical setting for individualized treatment of BC in the future. However, it should be pointed out that this study had some limitations: First, since development and validation of the signature in our study was based on TCGA database, it has not been verified in large-scale clinical samples, which may lead to selection bias. Secondly, biochemical experiments such as quantitative real-time PCR, immunohistochemistry, and flow cytometry must be designed to authenticate our model and further clarify the mechanism by which necroptosis-related lncRNAs regulate the pathological process of BC.