In the field of epigenetics research, human endogenous retrovirus (ERV) sequences are considered to play a crucial role in regulating gene expression and preserving genome stability. HERV-K elements can influence the expression levels of neighboring genes through diverse regulatory mechanisms, such as DNA methylation, histone modifications, and non-coding RNA interactions. These regulatory processes play critical roles in tumor occurrence and development. Understanding the impact of HERV-K elements on nearby gene expression is crucial for unraveling their potential implications in cancer biology and disease progression [12, 33, 34]. Here, we report for the first time the regulatory potential of 91 HERV-K proviruses on neighboring genes following abnormal activation in breast cancer. HERV-K may mediate the transcription of neighboring genes through proximal regulatory functions. HML-2 is considered the most recently integrated family of HERV-K elements, is the most biologically active subgroup within the HERV-K family, and possesses distinctive features in the human genome [14, 35], which provides a good system for exploring the regulatory function of HERV-K in neighboring genomic regions.

Constantly emerging evidence indicates that the genomic coordinates of HERV-K and sequence information of different insertion sites are being explored in the human genome. HERV-K sequences exist in two main forms in the human genome: one with a complete or nearly complete proviral sequence and the other composed only of LTR sequences [36]. Some insertion loci are referred to as proviral insertion sites because they contain one or more coding sequences for viral proteins. Indeed, genes located in the 0–35 kp region adjacent to HERV-K sequences in the human genome are known to be relatively densely distributed. One intriguing possibility is that the provirus, which is viral DNA integrated into the host genome, might undergo gene fusion with its neighboring host gene. This fusion event might have resulted in the creation of novel chimeric genes with potentially altered functions. Further research in this area could shed more light on the functional consequences of these potential gene fusions and their implications for human biology and health [37]. A total of 34 candidate genes were selected and screened to predict the prognosis of breast cancer effectively (p = 0.004). These genes were carefully chosen based on their potential association with HERV-K expression and involvement in immune-related functions. The goal was to identify genes that could serve as reliable prognostic indicators of breast cancer, allowing for a more accurate prediction of patient outcomes and disease progression.

Through rigorous analysis and validation, we aimed to determine the significance of 34 candidate genes in breast cancer prognosis and patient survival. CD48, SLAMF7, IGLL1, SLAMF1, IGHG2 and LRRC8A, deserve more investigations. These six genes were significantly enriched in immune function-related pathways. Among the 34 candidate genes, CD48 showed the highest expression (logFC = 15.52), followed by SLAMF1 (logFC = 13.35). All six key genes were identified as candidate genes in two or more samples. IGHG2 exhibited high expression in estrogen receptor-positive) and HER2 + (and human epidermal growth factor receptor 2-positive) breast cancer. On the other hand, CD48, SLAMF1, and SLAMF7 were highly expressed in HER2 + and triple-negative breast cancer (TNBC), and IGLL1 was highly expressed in ER + and TNBC tissues; These six key genes are closely associated with immune infiltration in breast cancer. In breast cancer, the high expression of SLAMF1, SLAMF7, CD48, and IGLL1 is positively correlated with the infiltration of B cells, CD8 + T cells, CD4 + T cells, macrophages, neutrophils, and dendritic cells. Conversely, the expression of LRRC8A is negatively correlated with the infiltration of the aforementioned immune cells. This suggests that the high expression of SLAMF1, SLAMF7, CD48, and IGLL1, as well as the low expression of LRRC8A, may induce or participate in activating the immune response within breast cancer tissue, regulating the activity and quantity of immune cells, thereby influencing the recognition and attack of tumor cells by immune cells. This infiltration of immune cells may be associated with the malignancy and prognosis of the tumor. IGLL1 showed a positive correlation with these immune cells, but its expression level was low, and the correlation was weak. However, the involvement of this gene in immune escape remains unclear. Further research is needed to investigate these results, and the results of the drug sensitivity analysis indicated that IGLL1, SLAMF7, SLAMF1, CD48, and LRRC8A might decrease sensitivity to specific chemotherapy drugs, further affecting the treatment efficacy for breast cancer. These variations in gene expression levels may influence the selection of treatment strategies and the prediction of therapeutic effectiveness. Further research is needed to elucidate the underlying mechanisms and potential clinical implications of this combination. Through in vitro experiments, we have confirmed that the activation of HERV-K in different subtypes of breast cancer cells (MCF-7, AU565, MDA-MB-231, and MB468) has a regulatory effect on neighboring genes. We observed certain differences in the expression of HERV-K_1q23.3, HERV-K_22q11.23, HERV-K_9q34.11, and HERV-K_14q32.33 and their adjacent genes in different breast cancer cell lines, which may be related to the heterogeneity of breast cancer cells [38, 39]. Furthermore, no differential expression was observed in HERV-K_14q32.33 across the four cell lines. However, a comprehensive analysis of gene expression results clearly indicates significant differential expression of HERV-K_1q23.3, SLAMF1, IGHG2, HERV-K_22q11.23, IGLL1, HERV-K_9q34.11, and LRRC8A in breast cancer cells, with a significant correlation between HERV-K and its corresponding neighboring genes. This finding is consistent with the trends identified in previous bioinformatics analyses.

The SLAM family (SLAMF) is a group of cell surface receptors that are involved in co-stimulation, cytokine production, and cytotoxicity and are crucial for regulating immune responses and facilitating communication between different immune cells [40,41,42]. The SLAM family (SLAMF) consists of nine cell surface receptors: CD150 (SLAM, SLAMF1), CD48 (SLAMF2, BLAST-1), CD229 (SLAMF3, Ly9), CD244 (SLAMF4, 2B4), CD84 (SLAMF5), CD352 (SLAMF6, Ly108, NTB-A), CD319 (SLAMF7, CRACC), CD353 (SLAMF8, BLAME), and CD84H (SLAMF9, SF2001). These receptors are involved in various aspects of immune function, such as co-stimulation, cytokine production, and cytotoxicity. Among the SLAM family members, CD48 (SLAMF2 and BLAST-1) plays a primary role in the adhesion and activation of immune cells. It is expressed on the surfaces of different immune cells and is crucial for mediating interactions between these cells. CD48 is involved in immune cell signaling and regulation, and its functions are essential for coordinating immune responses and promoting effective immune cell communication [43]. In the immune system, CD48 was the first discovered receptor for growth differentiation factor 15 (GDF15), and studies have shown that upregulated expression of GDF15 in HCC can regulate the suppressive function of natural Tregs (nTregs) through interaction with the CD48 receptor on T cells and transcriptional gene silencing mechanisms [44]. SLAMF1 is a co-stimulatory molecule involved in immune regulation that participates in host innate and adaptive responses [45]. In cancers like Hodgkin's lymphoma and chronic lymphocytic leukemia, IGLL1 also plays a crucial role in regulating the tumor microenvironment and determining the fate of malignant cells. Its involvement in these processes renders it a potential target for cancer treatment and further research [46]. SLAMF7 (also known as CD319, CRACC, and CS1) plays a central role in highly activated macrophage-related inflammatory diseases [36, 40]. SLAMF7 activation in inflammatory macrophages is a key pathway driving the pathology of acute and chronic inflammatory human diseases [47]. IGLL1, also known as immunoglobulin lambda-like polypeptide 1, is a member of the immunoglobulin gene superfamily that plays a vital role in B cell development [48]. IGHG2 (immunoglobulin heavy constant γ2) is a protein-coding gene that is associated with certain diseases such as immunoglobulin kappa light chain deficiency. Related pathways involve the production of C4 and C2 activators and the innate immune system [49]. Leucine-rich repeat protein A (LRRC8A), also known as SWELL1, is a core component of anionic channels (VRAC) [50]. LRRC8A is closely associated with the occurrence of multiple tumors [51,52,53].

An imbalance in the tumor immune microenvironment (TME) is one of the most significant characteristics of tumors [54]. The TME contains a variety of cell types, including tumor cells, stromal cells, and immune cells (T cells, B cells, and macrophages) [55]. The adaptive immune response mediated by immune cells plays a key role in tumor progression [56]. In various cancer types, the infiltration of immune cell populations has shown diverse prognostic outcomes, with immune cell types such as CD8 + T cells, B cells, CD4 + T cells, and neutrophil-macrophage dendritic cells playing crucial roles in the progression of specific tumors and influencing the response to immunotherapy [57,58,59]. In a small number of tumors, the infiltration of innate immune-related cells, such as natural killer (NK) cells, bone marrow-derived suppressor cells (MDSC), and DCs, is associated with prognosis, but the difference is great. Macrophage infiltration is a hallmark of solid cancers, and overall macrophage infiltration is associated with lower patient survival and treatment resistance [59]. The M1-type macrophages and M2-type macrophages are two distinct polarized subtypes of macrophages. M1-type macrophages are typically associated with antitumor immune responses and inflammation. Conversely, M2-type macrophages are associated with anti-inflammatory responses, immune regulation, and tumor growth[60]. The development of tumors is often associated with an imbalance between M1/M2-type macrophages[61, 62]. HERV-K may regulate immune responses, tumor progression, and treatment outcomes in breast cancer by potentially influencing macrophage activity and phenotype. Additionally, neutrophil mast cells and eosinophils have been associated with several tumor outcomes, with high neutrophil infiltration predicting poor prognosis and mast cells and eosinophils predicting good prognosis[63]. Therefore, we propose that the expression of immune-related genes in the breast cancer microenvironment is closely associated with immune cell infiltration. HERV-K_1q23.3, HERV-K_14q32.33, HERV-K_22q11.23, and HERV-K_9q34.11 may be markers of immune infiltration and poor prognosis in breast cancer.

Interestingly, the HERV-K sequence at locus 1q23.3, was completely integrated within the exon of CD48, with a 99.98% overlap with the CD48 gene sequence. The protovirus is 9232 bp in length, and with 5′LTR-gag-pol-env-3′LTR structure, we speculate that the change of CD48 expression is closely related to the insertion of HERVK_1q23.3 into the CD48 sequence. In addition, CD48, SLAMF7, IGLL1, SLAMF1, IGHG2, and LRRC8A showed different degrees of mutation. This may be another way in which HERV-K in the human genome shapes host genes [64].

Here, we found that in breast cancer samples, abnormally activated HERV-K viruses upregulated the expression of certain neighboring genes and downregulated the expression of certain genes. Of course, the ability to upregulate the expression of neighboring genes is more significant. For example, HERV-K_14q32.33 not only upregulates the expression of its upstream neighboring gene IGHG2 but also downregulates the expression of its downstream neighboring gene IGHAI. According to previous reports, HERV-K has a bidirectional promoter activity. We speculate that the potential antisense transcript at 14q32.33 may downregulate the expression of the IGHAI transcript [65]. These observations strongly support the hypothesis of diversity in HERV-regulated host gene pathways [12]. Further studies are needed to investigate how the original HERV-K viruses regulate their neighboring genes.

Our study extensively utilized publicly available raw data from nine laboratories to ensure high credibility and richness of the samples. These data substantially reflect the abnormal expression patterns of HERV-Ks in breast cancer, providing a substantial basis for further investigating the regulatory relationship between abnormal HERV-K expression and neighboring immune-related genes. However, it is essential to acknowledge the inherent limitations of bioinformatics analysis, despite our careful efforts and experimental validation. Moreover, whether our findings can be extrapolated to other types of tumors requires further validation and research.