Immune-related lncRNAs as predictors of survival in breast cancer: a prognostic signature

Among women, breast cancer is the most commonly diagnosed cancer and the leading cause of cancer-related death [1, 2]. Globally, it is reported that nearly 2.1 million new breast cancer cases were diagnosed, and more than 0.6 million women died of breast cancer in 2018 [3].Breast cancer is a highly heterogeneous disease and, thus, its etiology, pathological manifestations, and outcomes vary from person to person [4, 5]. Considering the high mortality and heterogeneity, it is urgent to identify suitable detection approaches for breast cancer prognosis biomarkers. Many studies have shown that some genes and mRNAs play significant roles as prognostic molecular markers in malignancies [6‐8]. Recently, research into tumor immunity has become the most rapidly advancing area within cancer. Immunotherapy provides the unprecedented opportunity to effectively treat malignancies owing to the essential involvement of the immune system in tumor development, progression, and therapy [9], especially in some malignancies such as hepatocellular carcinoma [10], early-stage squamous cell cancer of the anal canal [11], prostate cancer [12].

Long non-coding RNAs (lncRNAs) are classified as transcripts that are longer than 200 nucleotides and do not encode proteins. However, lncRNAs can physically interact with DNA, RNA, or protein. Through such interactions, lncRNAs are able to regulate gene expression at various levels such as transcriptional, post-transcriptional, and translational regulation.As a result, lncRNAs have important roles in the occurrence, progression, and prognosis of cancers and various other diseases [13‐15]. In addition, recent studies have indicated that lncRNAs have crucial functions in different phases of cancer immunity such as antigen presentation, immune activation, and immune cell infiltration [16, 17]. Thus, immune-related lncRNAs have drawn considerable attention. One study has reported the prognostic merit of an immune-related lncRNA signature in the prediction of overall survival (OS) in breast cancer [5]. However, this immune-related lncRNA signature has not been validated externally by other researches yet.

In the present study, we identified and validated a robust and reliable molecular signature for the prediction of survival in breast cancer patients. Our results validated a risk scoring model based on 8 immune-related lncRNAs. The model can be used as a reliable prognostic predictor, and the 8 lncRNAs could be potential therapeutic targets for breast cancer.

With the in-depth researches on lncRNAs and the immune system [21‐24], scholars have realized that immune-related lncRNAs may prove to be useful not only as potential prognostic biomarkers but could also provide novel therapeutic options. Nevertheless, the lack of validation cohorts is a limitation for the proper evaluation of the prognostic merit of potential biomarkers. In the current study, we identified a novel immune-related lncRNA molecular signature using Cox and Lasso regression analyses. The signature was then validated in a testing group and a total group, indicative of its robustness and reliability. The signature demonstrated good predictive performance and could effectively classify breast cancer patients into a high-risk and a low-risk group within the training and validation sets. The low-risk group had a significantly longer overall survival compared with the high-risk group. Further, the signature proved to be an independent prognostic factor based on multivariate Cox analysis, which revealed the signature’s reproducibility and reliability for breast cancer prognosis.

Among the 8 candidate lncRNAs, AC136475.2, AL161646.1, LINC01871, and AP000442.2 had not been previously reported. OTUD6B-AS1 and LINC00578 were discovered as prognostic signatures in breast cancer for the first time. Although AL122010.1 and AC245297.3 had been previously reported as prognostic signatures in breast cancer [25], our research characterized their functional involvement as lncRNAs associated with tumor immunity. OTUD6B-AS1 is transcribed from the opposite strand of the OTUD6B gene, which is located on chromosome 8 in head-to-head orientation to OTUD6BAS1 [26]. It was reported that high OTUD6B-AS1 expression indicates poor prognosis in ovarian cancer [27]. However, Gang Wang et al. [28] found that high OTUD6B-AS1 expression was associated with improved survival and inhibited clear cell renal cell carcinoma proliferation via the Wnt/β-catenin signaling pathway. Zhuolu Wang et al. found that OTUD6B-AS1 inhibits viability, migration, and invasion of Thyroid Carcinoma by Targeting miR-183-5p and miR-21 [29]. Further more, it was reported that the expression of miR-21 in cells of the tumor immune infiltrate, and in particular in macrophages, was responsible for promoting tumor growth [30]. Activation-induced up regulation of miR-21 biases the transcriptome of differentiating T cells away from memory T cells and toward inflammatory effector T cells [31]. In the current study, we found that the overall survival was lower under high OTUD6B-AS1 expression. Further, expression was higher in the N1 and N2 groups when compared to the N0 group. We speculated that OTUD6B-AS1 affects macrophages and differentiating T cells through miR-21, then affects immune status, and finally affects the occurrence and development of tumor. Therefore, further research and mechanistic insight are required. LINC00578 was found as a potential biomarker in lung adenocarcinoma [32, 33], major depressive disorder [34], and pancreatic cancer [35]. Although the AC243960.1 and OIP5-AS1 lncRNAs were the most positively and negatively correlated, respectively, with immune genes, they did not appear in the signature. It was indicated that high AC243960.1 expression indicates better prognosis in breast cancer, however, no statistically significant interactions between OIP5-AS1 expression and OS was found (Additional file 7: Figure S5). This may be due to the deletion of some relevant data during regression analysis.

As breast cancer is a highly heterogeneous disease, scientists divide breast cancers into different clinically relevant molecular subtypes based on the expression levels of the estrogen receptor (ER), progesterone receptor (PR), and HER2 [36‐38]. Distinct prevalence, prognosis, and systemic therapies are utilized in the management of these different breast cancer subtypes [39‐41]. HER2-overexpressing (ER and PR− , HER2 +) and triple-negative (ER and PR− , HER2−) subtypes are known to be more aggressive and have poorer outcomes [38, 42]. Our results indicated that the expression of LINC01871 was high on both subtypes. Meanwhile, and LINC01871 shows a strong positive correlation with immune genes such as GZMB, CTLA4, PDCD1 etc. (Additional file 5: Figure S3). GZMB, the most potent cytotoxic molecules, act mainly as antitumoral and anti-infectious factors. However, when expressed by immune regulatory cells it may contribute to immune evasion of specific cancer types [43]. Cytotoxic T lymphocyte-associated antigen 4 (CTLA4) and programmed death-1 (PD-1; encoded by the PDCD1 gene) represent crucial immune checkpoints, the blockade of which can potentiate anti-tumour immunity [44, 45]. Therefore, it is suggested that LINC01871 may play an important role, particularly related to the above immune processes and immune genes, in the development of breast cancer in the two phenotypes, which requires in-depth investigation in the future.

Finally, GSEA further confirmed the robust connection of the signature with the immune response. Samples from patients with low-risk scores were associated with positive regulation of adaptive immune response. Further, patients with a high-risk score exhibited greater adaptive immune resistance. Adaptive immune resistance is a process during which cancers change their phenotype in response to a cytotoxic or pro-inflammatory immune response, thereby evading it [46]. Inhibition of adaptive immune resistance is the mechanistic foundation of responses to PD-1 or PD-L1 inhibition [47, 48], which has made a significant contribution to the treatment of breast cancer [49, 50]. Our prognostic signature may provide directions for predicting the efficacy or studying the mechanism of PD-1 or PD-L1 inhibition immunotherapy. Lymphocytes, including T cells, B cells, and natural killer cells, are the main force underlying immune defense mechanisms. For them to perform their immune function, lymphocytes must be activated either through the recognition and binding of antigens or through stimulation by cytokines [51‐53]. In this study, samples from patients with low-risk scores were associated with the positive regulation of lymphocyte activation, which may indicate that low-risk patients have a more active immune status and better immune defense than do high-risk patients. T cells have been considered as having a significant role in immune surveillance and tumor eradication. On the basis of this paradigm, over the past quarter century, T cell-based cancer therapies have achieved success in patients [54, 55].

The advantage of the current study was that our signature is based on population databases and high-throughput sequencing data. Further, both exploration and validation were used in order to evaluate the risk score method. Tumor immunology has become the most rapidly advancing area of cancer research, and immunotherapy has provided promising treatment in recent years. The current study utilized a new immune-related prognostic approach for breast cancer. However, there were some limitations in this study. First, as the analyzed data was obtained from online databases, the study was of retrospective nature. Second, there is no in vitro or in vivo experimental data confirming our findings. In addition, we did not explore the potential mechanisms of investigated lncRNAs. Thus, more functional studies on the 8 lncRNAs, alone and in combination, should be carried out to further test the predictive accuracy of the signature and discover potential immune-related mechanisms. Of note, we are currently working on the clinical validation and mechanistic elucidation of these results. In various studies, gene expression differences between cancer and normal tissues are compared for the screening of prognostic genes. This may leave out certain genes with little expression differences between cancer and normal tissues. Such genes with no obvious expression difference may have a great influence on the biological behavior of tumors, chemotherapy, immunotherapy, and other factors affecting the survival of patients. Despite these limitations, to the best of our knowledge, this study is the first to report the external validation of an established immune-related lncRNA signature for breast cancer. The eight immune-related lncRNAs had never been studied in breast cancer.

In conclusion, we have identified an 8 immune-related lncRNA signature as a potential prognostic tool for breast cancer patients. It is strongly connected to the risk value, tumor status, and OS. The signature provides a novel insight into immune-related lncRNAs in breast cancer and identifies potential biomarkers for prognosis and immunotherapy.