Breast cancer is the highest incidence and deadliest type of carcinoma for women worldwide, presenting highly heterogeneous biological and clinical characteristics. As a kind of nanosized vesicles, exosomes play vital roles in tumor development and progression. Importantly, they can regulate cell-to-cell communication in the tumor microenvironment via proteins, lipids and RNA cargo transferred. Most recently, ICB therapy has made promising progress in cancer immunotherapy. In 2017, Chen and colleagues raised the idea that tumors with lower PD-L1 expression level and fewer infiltrating cytotoxic T cells were considered as immune “cold” tumors, where ICB therapy achieved only limited results. Fatally, “cold” tumors account for the majority. Increasing evidence suggested that exosomes had the potentiality to work as biomarkers for a variety of malignant tumors including breast carcinoma. Previous study in pancreatic cancer identified that exosomes released from tumor contained a membrane bound protein called GPC1, considered as a sensitive and unique biomarker in early-stage disease. Interestingly, a study in breast cancer indicated that the expression level of serum exosomal-annexin A2 (exo-AnxA2) could be detected higher in women with carcinoma compared with non-cancer, particularly for triple-negative breast cancer (TNBC) rather than luminal or HER2-positive breast cancer. These results suggested that exosomes played an important role in BC progression and the possibility to construct a prognostic model by these exosome-related genes. In the current study, we adopted bioinformatics analysis to examine changes in the expression profiles of 121 exosome-related genes in breast cancer and the relationship with OS. Among them, 7 exosome-related genes were identified to establish a novel prognostic signature. Subsequently, the BC patients in TCGA and GEO cohorts were divided into high and low risk groups respectively according to the prognostic signature. Moreover, the DEGs between the high-risk group and the low-risk group were identified and the functional analysis were further performed, proposing that immune-related biological processes were highly enriched. Eventually, the infiltration ratios of distinct immune cells in breast cancer samples were analyzed. The results revealed that the groups at high risk of exosomes were immunologically 'cold', while the groups at low risk were immunologically 'hot'.
The 7 prognostic genes associated with exosomes consist of EpCAM, PIGR, KRT14, DOK7, CD24, CYP19A1, and CXCL13. As one of the first tumor-related antigens, epithelial cell adhesion/activating molecule (EpCAM/CD326) has been reported highly overexpressed in primary and metastatic breast cancer, leading to poor prognosis. Furthermore, study in hepatocellular carcinoma (HCC) suggested that tumor growth and invasion were associated with EpCAM-positive cells, which was one of the components of targeting Wnt/beta-catenin signaling pathway. Cluster of differentiation 24 (CD24) is a glycosyl-phosphatidyl-inositol (GPI)-anchored glycoprotein, that has been demonstrated as a vital role in multiple areas. In cancer, CD24 is highly expressed in various tumor cells, including breast cancer cells, and associated with the growth, invasion, and migration of tumor cells[36–40]. In immunology, as a primarily costimulatory molecule, CD24 can achieve effective immunosuppression and tumor immune escape via activating a series of intracellular signal pathways and regulating multiple immune cells, for instance, T cells, B cells, macrophages, and NK cells. Notably, Barkal and colleagues revealed that the macrophages enhanced their capability to engulf tumors and slowed down the development of macrophage-dependent tumors in vivo through gene knockout to blocking CD24 and Siglec-10. This is consistent with our conclusion, the infiltration level of M0 macrophages is higher in high-risk group. Polymeric immunoglobulin receptor (PIGR) is a critical element of the mucosal immune system and intermediates epithelial cell transfection of immunoglobulins. And the expression level of PIGR was demonstrated decreased in nasopharyngeal carcinoma cells by Qi and colleagues, related to poor prognosis[43, 44]. Several studies suggested DOK7 as a potential tumor-suppressor gene, because of the significantly low expression levels in BC tissues compared with normal tissues. The lower expression level of DOK7 was associated with the greater aggressive clinical behaviors and poorer prognosis in BC. The mechanistic studies by Yue et al. illustrated that DOK7 inhibited proliferation and invasion of BC cells through PI3K/PTEN/AKT pathway. CXCL13 may be another potential protective factor for BC patients’ prognosis. Although, CXCL13 has been reported in many types of carcinomas to drive signaling pathways associated with proliferation and invasive, including PI3K/AKT pathway and Wnt/beta-catenin signaling pathway, in patients with HER2-positive BC or TNBC, increased CXCL13 corresponded with better survival. In addition, CXCL13 has been shown to increase B cell and T cell infiltration in multiple tumor types and associate with greater prognosis and survival[50, 51].
In our study, the results of functional enrichment analysis indicated that many GO terms and KEGG pathways were associated with biological processes and immune cells activation. In 1996, immunologists were interested in the relationship between exosomes and cells from immune system. They found that Epstein-Barr virus-transformed B lymphocytes could secrete exosomes by fusion of MVBs with the plasma membrane. Furthermore, some discoveries suggested that exosomes played key roles in adaptive immune responses. Exosomes released from these kinds of cells harbor MHC class II dimers bound to antigenic peptides and the exosomes were indicated to present the MHC–peptide complexes to specific T cells. In addition, dendritic cells (DCs) in mice were reported to secrete exosomes with functional MHC class I–peptide complexes, which could improve the triggering of CD8+ T-lymphocyte-dependent immune responses[53, 54]. Recently, exosomes have been shown to be involved in promoting immune responses. Studies in human pancreatic and colorectal tumours conducted by Gastpar R and his colleagues suggested that NK cells cultured with tumour-derived Hsp70-positive exosomes were induced to liberate granzyme B that activated apoptosis. As the mechanisms of exosomes in antigen-specific immune responses were better understood, several studies showed that tumor-released exosomes also carried a variety of immunosuppressive molecules, such as CD8 and CD4 T lymphocytes[56, 57], NK cells, regulatory T lymphocytes and myeloid cells. Hence, we reasonably assumed that exosomes were closely associated with anti-tumor immunity in BC. Furtherly, The CIBERSORT algorithm was applied to derive the ratios of various kinds of tumor-infiltrating immune cells to reveal the relationship between exosomes and immune cell infiltration in BC. The result indicated that the infiltrated tumor-killing immune cells were significantly reduced in high-risk group instead of low-risk group, such as CD8+ T cells and activated NK cells. However, the immune cells that promoted tumor progression and migration, M2 macrophages and M0 macrophages, were reduced in low-risk group instead of high-risk group. As a marker of tumor progression and drug resistance, TIME was characterized by tumor inflammation promoting and tumor cells immune surveillance. Recent studies in tumors observed that some immune activities were associated with exosomes. Studies in exosomes secreted from prostate cancer cells indicated that exosomes contained ligands for natural killer group 2D (NKG2D) could downregulate the expression of NKG2D on NK cells and impair the cytotoxicity of NK cells[58, 62]. Besides, those exosomes secreted by HCC cells with chemotherapy treatment efficiently stimulated NK cells to product the granzyme B, hence promoted the tumoricidal function. In tumor microenvironment, those tumor cell-derived exosomes could influence the DCs status. The results of experiments in vitro indicated that exosomes produced by TS/A BC cells could block the differentiation progression of DC from myeloid cells. In terms of macrophages, type 1 macrophages (M1) were involved in anti-tumor immune responses by acting as antigen-presenting cells whereas type 2 macrophages (M2) acted in the pro-tumor immunity as the common phenotype of tumor-related macrophages. Several studies revealed that the M2 status was correlated with the progression of tumors and poor prognostic outcomes of patients. Yang M et al. reported that exosomes derived from M2 could promote BC cell growth and invasion by transferring miR-223. Additionally, neutrophils status was reported to be regulated and differentiated via the stimulation from TME, as a result, to further moderate tumor immune responses and regulate tumor progression. Bobrie A and colleagues investigated on the value to Rab27a in exosome secreted by BC cells, consequently, exosomes were identified to induce systemic mobilization to neutrophils to facilitate tumor progression. Not only different kinds of immune cells, but also other cellular components of TME, including MSCs, endothelial cells and fibroblasts, played significant roles in tumor progression. Accordingly, a conclusion was raised that the exosomes were markedly associated with the ratio of tumor-related immune cells infiltrating in BC, and the low-risk cohort tended to possess higher ratio of cytotoxic lymphocyte infiltration.
Recent research findings suggested that exosomes were able to control the core immunologic processes and regulate inflammatory response. Immune cell exosomes were identified to involve in stem cell mobilization, immunological regulation, and tissue remodeling. Immunotherapy is increasingly becoming the key to BC treatment. Yang Y et al. found that PD-L1 could be carried by BC cell-derived exosomes and transferred to tumor cells expressed low levels of PD-L1 to blocked T cell activity. Notably, exosome-related immunotherapeutic exoPDL1 was able to be applied to develop novel drugs with minimum toxicity and considerable clinical effectiveness. Furthermore, immune checkpoints, a key of facilitating tumor immunosuppression, were analyzed in different risk groups to detect the correlation between exosomes and immune checkpoint inhibitors. Stimulation of immune checkpoint targets can block tumor attack. In our study, high-risk group predicated on 7-exosome-related gene risk model was found to associate with lower expression levels of immune checkpoint genes and poor clinical outcomes. As a result, mammary tumor in high-risk group were considered immunologically “cold” and difficult to gain benefits from ICBs. However, the tumors in low-risk group were considered immunologically “hot” and more possible to get benefits from ICBs. Accordingly, the exosome-related risk model had highly potential to predict the efficacy of ICBs in treatment for patients suffered from BC.
It was the first study to construct an exosome-related risk model for breast cancer based on 7 exosome-related genes by use of public databases with retrospective data. This risk model can serve as an independent prognostic factor in BC patients. In addition, some limitations in current research should be noticed. Some of the clinical data in the TCGA or GEO cohorts were incomplete and the absent data may not be random, leading to the bias in the analysis of clinical relevance. Therefore, it requires more extensive multi-center clinical validation to further back up our ideas. Besides, this prognostic model was built only from exosome-related genes and various other hotspot biomarkers were not enrolled. Consequently, further experimental validation of the relationship between immune cells and exosomes is needed to provide new perspectives in immunotherapy and tumor treatment.