Role of SERPINA1 in the tumor immune microenvironment of breast cancer and construction of a prognostic model

Abstract

Background

Breast cancer is a common malignancy where the immune system plays a key role in disease progression and treatment response. SERPINA1 is an immune-related gene, but its function in breast cancer remains unclear. This study explored the expression, immune role, and prognostic value of SERPINA1 in breast cancer.

Methods

We analyzed data from TCGA and other public databases to evaluate SERPINA1 expression and performed Western blot and qPCR experiments for validation. Various bioinformatics tools were used to assess immune-related functions, identify differentially expressed genes, conduct pathway enrichment analyses, and construct a prognostic risk model. External datasets were used for validation, and immune activity and drug sensitivity were compared between different risk groups.

Results

SERPINA1 was significantly overexpressed in breast cancer tissues compared to normal tissues, and high expression was associated with better overall survival. High SERPINA1 expression correlated with enhanced activation of immune-related pathways, such as T and B-cell signaling. Patients with high SERPINA1 levels showed higher immune and stromal scores, increased infiltration of CD8⁺ T cells, macrophages, and dendritic cells, and elevated expression of immune checkpoint molecules. A prognostic model based on SERPINA1-associated immune genes effectively stratified patient survival risk and was validated in multiple external datasets. Patients in the low-risk group had stronger immune activity and were more sensitive to various chemotherapy and targeted drugs, suggesting the model’s potential in guiding personalized treatment.

Conclusion

SERPINA1 is closely linked to favorable prognosis and an active immune microenvironment in breast cancer. The SERPINA1-based prognostic model may be useful for survival prediction and personalized immunotherapy strategies.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-026-04555-7.

Introduction

On April 11, 2025, according to the latest data released by the International Agency for Research on Cancer (IARC) under the World Health Organization (WHO), breast cancer (BRCA) has become the second most commonly diagnosed cancer globally, with 2.3 million new cases. With 670,000 deaths, it has surpassed gastric cancer to become the fourth leading cause of cancer-related mortality worldwide, far exceeding other gynecologic malignancies such as cervical and ovarian cancers (Global Cancer Observatory: Cancer Today. Available from: https://gco.iarc.who.int/today, accessed [11/04/2025]). BRCA now accounts for 24.5% of all cancers in women [1, 2]. Although early-stage BRCA can achieve favorable prognosis through surgery combined with radiotherapy and chemotherapy, the 5-year survival rate for patients with advanced or metastatic disease remains below 30% [3]. Cancer immunotherapy, particularly immune checkpoint inhibitors (ICIs), has revolutionized treatment paradigms recently. However, the response rate of BRCA to immunotherapy is significantly lower than that observed in melanoma or non-small cell lung cancer [4]. Studies have identified the heterogeneity and immunosuppressive nature of the tumor immune microenvironment (TIME) as key barriers to therapeutic efficacy [5]. Therefore, elucidating the dynamic regulatory networks of TIME in BRCA and identifying actionable molecular targets represent critical directions for overcoming current therapeutic limitations. Serpin Family A Member 1 (SERPINA1), also known as α1-antitrypsin (AAT), is an acute-phase protein synthesized primarily in the liver. It modulates inflammatory responses by inhibiting serine proteases such as neutrophil elastase and proteinase G [6]. While early studies mainly focused on its protective role in chronic obstructive pulmonary disease and liver fibrosis [7], recent research has revealed a more complex involvement of SERPINA1 in tumor progression and immune evasion.

For instance, in pancreatic cancer, SERPINA1 promotes the activation of cancer-associated fibroblasts (CAFs) by stimulating the IL-6/STAT3 signaling pathway, thereby remodeling an immunosuppressive microenvironment [8]. In colorectal cancer, elevated SERPINA1 expression has been significantly associated with increased infiltration of regulatory T cells (Tregs) and suppressed function of cytotoxic T lymphocytes (CTLs) [9–11].

Notably, the immunomodulatory role of SERPINA1 appears to be highly tumor type–dependent. In lung cancer, its overexpression impairs tumor-specific T cell responses by inhibiting the proteolytic activity of antigen-presenting cells (APCs) [12]. In contrast, deleting the SERPINA1 gene in melanoma has been linked to enhanced tumor metastasis [13]. These seemingly contradictory findings highlight the urgent need for a systematic investigation into the mechanistic roles of SERPINA1 within the tumor microenvironment, particularly in BRCA, where its functional implications remain poorly understood. Currently, research on the association between SERPINA1 and BRCA remains highly limited. Early clinical studies have reported significantly elevated serum SERPINA1 levels in BRCA patients; however, its correlation with tumor stage or prognosis remains controversial [12, 14]. A cohort study focusing on hormone receptor–positive BRCA revealed a positive association between SERPINA1 mRNA expression and activation of the estrogen receptor (ER) signaling pathway [15]. In contrast, another study suggested that SERPINA1 may regulate BRCA cell invasiveness by modulation of matrix metalloproteinases (MMPs) [16, 17]. Nevertheless, these studies have not explored the interactions between SERPINA1 and the TIME in BRCA nor conducted systematic validations based on large-scale multi-omics datasets. Although numerous studies have confirmed the close relationship between immune cell infiltration and BRCA prognosis [18] and key regulatory molecules such as PD-L1 [18] and CXCL9 [19] have been extensively investigated, whether SERPINA1 contributes to BRCA progression by modulating immune cell function remains largely unknown.

Notably, the immunomodulatory role of SERPINA1 is highly tumor-type dependent: it exerts immunosuppressive effects in pancreatic and colorectal cancer (e.g., activating CAFs via IL-6/STAT3, increasing Tregs to suppress CTLs) [8, 9], yet its function in breast cancer (BRCA) remains poorly defined. Early BRCA-related studies on SERPINA1 were limited to serum level detection, estrogen receptor signaling association, or MMP regulation [14–16], lacking systematic multi-omics validation and in-depth analysis of the tumor immune microenvironment (TIME). Given the unique heterogeneity of the BRCA TIME—characterized by higher pro-inflammatory cytokine levels and distinct immune cell composition compared to other malignancies [5, 35]—direct extrapolation of findings from other cancers is inappropriate. Thus, systematic investigation into SERPINA1’s expression, prognostic value, and immunoregulatory mechanisms in BRCA is urgently needed to fill this knowledge gap.

In this study, we systematically analyzed the expression patterns of SERPINA1 in BRCA and its associations with patient prognosis, the TIME, and therapeutic responses. By integrating data from TCGA and other public databases with experimental validation, we found that high SERPINA1 expression was strongly associated with the activation of multiple immune-related pathways, including T cell receptor signaling, B cell receptor signaling, and NK cell–mediated cytotoxicity. This suggests that SERPINA1 may contribute to improved prognosis by modulating antitumor immune responses.

Further analysis of the cancer-immunity cycle revealed that although most immune steps were more active in the high-expression group, functional limitations were observed during the tumor cell elimination phase, potentially reflecting the involvement of immunosuppressive mechanisms. An immune-related risk score model based on SERPINA1 expression effectively stratified patients by immune status and predicted sensitivity to various chemotherapeutic and targeted agents.

These findings suggest that SERPINA1 may be a potential target for immunotherapy or combination treatment strategies in BRCA. We hope this study will contribute to clinical decision-making and provide new insights into therapeutic approaches.

Methods

Figure 1 shows the workflow of this study.

Fig. 1.

Workflow of this study. A Data sources; B Expression characteristics of SERPINA1 in BRCA and its potential functional analysis; C Correlations between SERPINA1 and the immunological status in TME in BRCA; D Correlation analysis between SERPINA1 expression levels and characteristics of the TIME in BRCA; E Lasso regression for prognostic model construction and validation.; G Nomogram, immune, and drug sensitivity analysis

Data acquisition and preprocessing

RNA sequencing data, clinical information, and somatic mutation profiles of BRCA patients were obtained from the TCGA-BRCA project (https://portal.gdc.cancer.gov/, accessed on March 15, 2025). The dataset included 1,113 tumor tissue samples and 113 adjacent normal tissue samples, of which 937 cases had complete survival information. External validation datasets GSE20685 and GSE42568 were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/).

To validate the database-derived findings, paired tumor and adjacent normal breast tissue samples were collected from four BRCA patients treated at the Department of Breast and Thyroid Surgery, Yiyang Central Hospital, following approval by the institutional ethics committee (Ethical review No. (2022) Hospital ethics review No.034, Project ID: [2022]NO.108). SERPINA1 protein expression was assessed by Western blot using a specific antibody (PROTEINTECH, #66,135–1-Ig, 1:8000), with β-actin as the internal control (PROTEINTECH, #60,009–1-Ig, 1:10,000). The raw data of the SERPINA1 WB analysis is in the S1 File.

Quantitative PCR (qPCR) was performed using SYBR Green Master Mix (TransGen Biotech, #AQ601-04). The raw data for qPCR and the primer sequences are provided in the S File.

Analysis of SERPINA1 expression levels in BRCA

The GEPIA2 online tool (http://gepia2.cancer-pku.cn) was used to compare SERPINA1 expression levels between BRCA and normal breast tissues. The thresholds for statistical significance were set at |log₂ fold change (FC)|> 1 and false discovery rate (FDR) < 0.05. To validate the protein-level expression pattern of SERPINA1 in BRCA tissues, the HPA database (https://www.proteinatlas.org/) was queried for immunohistochemistry (IHC) images and subcellular localization data of SERPINA1 in BRCA and normal tissues[20].

Mutation profiling and functional pathway analysis

Somatic mutation data from TCGA were processed and visualized using the "maftools" package to compare the frequency and types of mutations between groups stratified by SERPINA1 expression and risk scores.

To explore the molecular characteristics associated with different SERPINA1 expression levels, Gene Set Variation Analysis (GSVA) was performed using the KEGG gene set collection (c2.cp.kegg.v7.4)[21].

Tumor immune microenvironment analysis

The immune landscape of the BRCA tumor microenvironment (TME) was evaluated across four dimensions in this study: (1) the expression profiles of immune regulatory molecules, (2) the activation status of cancer immunity cycle processes, (3) the abundance of tumor-infiltrating immune cells (TIICs), and (4) the differential expression of immune checkpoint molecules associated with immunosuppression.

122 immune-related regulatory factors—including MHC molecules, chemokines, immunostimulatory molecules, and immune receptors—were integrated for a systematic transcriptional analysis of immune characteristics in BRCA (S3 File. Table S1) [22].

To comprehensively capture the antitumor immune response, we adopted the framework of the “cancer-immunity cycle,” which divides the immune response into seven key steps: release of tumor antigens, antigen presentation, priming and activation of T cells, recruitment and infiltration of immune cells, recognition of tumor cells, and eventual tumor cell elimination [23]. We applied the single-sample gene set enrichment analysis (ssGSEA) method to score the immune activity of each step in the cancer-immunity cycle based on gene expression profiles from individual BRCA samples (S3 File. Table S2).

To comprehensively assess immune cell infiltration, six distinct deconvolution algorithms were integrated into this study, including TIMER, CIBERSORT, quanTIseq, MCP, TISIDB, and TIP, ensuring both the robustness and consistency of immune infiltration estimates (S3 File. Table S3)[24–29].

In addition, based on previously reported immune effector genes, we further evaluated the functional activity of specific immune cell subtypes (S3 File. Table S4). To explore the expression characteristics of immunosuppressive regulators within the TME, we also analyzed a curated panel of 22 immune checkpoint–related genes with potential relevance to immunotherapy [30] and compared their expression across different subgroups.

Construction and validation of the risk model

Based on the SERPINA1 high vs. low expression groups and immune score stratification, differential expression analysis was performed using the ‘DESeq2’ package to identify DEGs, with screening thresholds set as |log₂FC|> 1 and FDR < 0.05. A total of 803 DEGs were obtained, and univariate Cox regression analysis was further performed to screen these DEGs for survival-associated genes (n = 453). Subsequently, the glmnet package was used to construct a prognostic risk model via least absolute shrinkage and selection operator (LASSO) regression, with the optimal λ value determined as 0.0073 through tenfold cross-validation. A seven-gene risk score model was established accordingly. These genes were not only statistically prioritized but also functionally consistent with SERPINA1’s immunoregulatory role, supported by existing literature: RAC2 regulates immune cell activation and migration [31], a key step in antitumor immunity; APOBEC3D mediates tumor immune editing and correlates with immunotherapy response [32]; FAM159A modulates inflammatory responses [33]; TFPI2 participates in immune microenvironment remodeling [34]; ADRB1 regulates immune cell recruitment [35]; IGHD is essential for B cell maturation [36]; PIGR enhances mucosal immunity and immune surveillance [37]. Their established roles in tumor immunity and prognosis validate their rational inclusion in the SERPINA1-associated model.

The risk score was calculated using the following formula: Risk Score = ∑ (Expression_i × Coefficient_i).

The model's predictive performance was evaluated using Kaplan–Meier survival analysis and the “timeROC “ package [38]. To assess the risk score's clinical utility further, we integrated it with conventional clinical variables to construct a nomogram for BRCA patients. The accuracy and predictive efficacy of the nomogram were validated using calibration curves and ROC curve analysis.

Drug sensitivity prediction

Drug sensitivity to 198 chemotherapeutic and targeted agents was predicted for patients in the high- and low-risk groups using the “oncoPredict “ package [39]. Differences in drug sensitivity between the two risk groups were assessed using the Wilcoxon rank-sum test.

Single-cell RNA-seq data analysis

To validate the cellular source of SERPINA1 and its tumor-specific overexpression pattern, single-cell RNA-seq datasets of BRCA and paired adjacent normal tissues (GSE176078, GSE161529) were downloaded from the GEO database.

Raw data processing and quality control were performed using the “Seurat” package (v4.3.0) in R (v4.3.2): low-quality cells (mitochondrial gene content > 10%, ribosomal gene content > 60%, unique molecular identifiers (UMIs) < 200 or > 6000) and doublets were filtered out. Normalization was conducted via the LogNormalize method, and highly variable genes were identified to support subsequent dimensionality reduction and clustering.

Principal component analysis (PCA) and Uniform Manifold Approximation and Projection (UMAP) were applied for dimensionality reduction and visualization. Cell clustering was performed with the FindClusters function (resolution = 0.6), and cell type annotation was achieved using canonical marker genes.

SERPINA1 expression across cell subsets and between tumor/normal tissues was visualized using violin plots and UMAP plots. Tumor samples were stratified into SERPINA1-high and SERPINA1-low groups based on median expression. Differential expression analysis (DEGs) was performed for fibroblasts, basal cells, and macrophages using the FindMarkers function (|avg_log2FC|> 0.25, pvalue < 0.05), and KEGG pathway enrichment analysis of DEGs was conducted with the clusterProfiler package.

Statistical analysis

All statistical analyses were conducted in the R environment (version 4.3.2). Differences between the two groups were assessed using the Wilcoxon rank-sum test. Correlation analyses were performed using Spearman’s method. Kaplan–Meier survival analysis was conducted using the log-rank test. LASSO regression was applied for variable selection, and ROC curves were used to evaluate model performance. A P value < 0.05 was considered statistically significant.

Results

Expression profile, prognostic significance, and immunoregulatory function of SERPINA1 in BRCA

Using GEPIA2 to perform differential expression analysis on TCGA BRCA (BRCA) samples, we found that SERPINA1 expression was significantly higher in tumor tissues compared to normal breast tissues (Fig. 2A). To validate this database-derived finding, we conducted Western blot (WB) and quantitative PCR (qPCR) analyses on paired BRCA and adjacent normal tissues. Both assays consistently demonstrated a marked upregulation of SERPINA1 in tumor samples (Fig. 2B, C). We further examined the expression pattern and localization of SERPINA1 across different histological subtypes of BRCA. Immunohistochemistry images from the Human Protein Atlas (HPA) revealed high expression of SERPINA1 in both ductal carcinoma and lobular carcinoma, with protein localization primarily in the cytoplasm and cell membrane (Fig. 2D).

Fig. 2.

Expression characteristics of SERPINA1 in BRCA and its potential functional analysis. A Box plot showing the difference in SERPINA1 expression levels between BRCA and normal tissues (Tumor = 1085; Normal = 291. (*p < 0.05) B Western blot analysis of SERPINA1 protein expression levels in BRCA tumor tissues and normal tissues (T = Tumor, N = Normal). (***p < 0.001) C qPCR validation of SERPINA1 mRNA expression in tumor and normal tissues. (****p < 0.001) D Immunohistochemical localization of SERPINA1 in BRCA tissues from the HPA (Patient ID: 2805 is Lobular carcinoma; Patient ID: 1910 is Duct carcinoma). E Kaplan–Meier survival curve analysis of SERPINA1 expression levels and overall survival (OS) in BRCA. F GSVA of TCGA-BRCA patients. (*p < 0.001)

To assess its clinical significance, Kaplan–Meier analysis using TCGA survival data indicated that patients with high SERPINA1 expression had significantly better OS compared to those with low expression (Fig. 2E). However, stratified analysis by clinical features revealed no significant differences in SERPINA1 expression across subgroups defined by age, clinical stage, tumor size, lymph node status, distant metastasis, or PAM50 molecular subtype (Supplementary Fig. 1A), suggesting that the prognostic value of SERPINA1 may be independent of conventional clinicopathological factors.

To further investigate the biological mechanisms underlying differential SERPINA1 expression, we first compared the somatic mutation profiles between the high and low-expression groups. The analysis revealed high mutation rates in both groups—87.69% in the high-expression group and 84.03% in the low-expression group. PIK3CA and TP53 were the most frequently mutated genes in both groups, with slightly higher mutation frequencies observed in the high-expression group (39% and 31%, respectively). Although the two groups had some differences in mutation types—such as missense mutations and frameshift insertions/deletions—the overall distribution of mutation profiles was largely similar. It did not exhibit statistically significant differences (Supplementary Fig. 1B).

Finally, to explore the potential biological functions associated with differential SERPINA1 expression, we performed GSVA for pathway enrichment. The results showed that BRCA patients with high SERPINA1 expression were significantly enriched in multiple immune-related signaling pathways, including the T cell receptor signaling pathway, B cell receptor signaling pathway, NK cell–mediated cytotoxicity, antigen processing and presentation, and cytokine–cytokine receptor interaction (Fig. 2F). These findings suggest that SERPINA1 may improve patient prognosis by enhancing antitumor immune response mechanisms.

Single-cell RNA-seq validation of SERPINA1’s cell-specific expression and tissue differences

To validate the cellular localization and tumor-specific overexpression of SERPINA1, we analyzed single-cell RNA-seq datasets of BRCA and adjacent normal tissues. Single-cell RNA-seq data were obtained from the study by Regner et al. [40]. After quality control and clustering, eight major cell types were identified in the BRCA microenvironment, including epithelial cells (EPCAM), fibroblasts (COL1A1), macrophages (CD68), T cells (CD3D), B cells (CD79A), dendritic cells (CD1C), basal cells (KRT5), and endothelial cells (VWF) (Fig. 3A).

Fig. 3.

Single-cell RNA-seq validation of SERPINA1 in BRCA. A Cell type annotation of BRCA microenvironment (UMAP plot). B Scatter-violin plot displaying SERPINA1 expression levels across different cell subsets (each dot represents a single cell; violin shape indicates expression distribution. C UMAP plot colored by SERPINA1 expression intensity (red-yellow: high expression; light blue: low expression), showing the spatial distribution of SERPINA1 in the BRCA microenvironment. D Violin plots comparing SERPINA1 expression between tumor and adjacent normal tissues across different cell subsets (****P < 0.0001, **P < 0.01; ns: not significant; paired t-test). E Volcano plots of differentially expressed genes (DEGs) in fibroblasts, basal cells, and macrophages between SERPINA1-high and SERPINA1-low groups (red: upregulated; blue: downregulated; |log₂FC|> 1, FDR < 0.05). F KEGG pathway enrichment analysis of DEGs in fibroblasts, basal cells, and macrophages (adjusted P < 0.05, top 2 pathways shown)

Violin plot analysis showed that SERPINA1 exhibited the highest expression in macrophages, followed by luminal cells, basal cells, and fibroblasts; expression levels were extremely low in T cells, B cells, plasmablasts, mast cells, and endothelial cells (Fig. 3B). UMAP visualization further confirmed this pattern: high SERPINA1 expression was concentrated in the macrophage, luminal cell, basal cell, and fibroblast clusters, while other cell clusters showed only low-level expression (Fig. 3C). Comparison between tumor and adjacent normal tissues revealed that SERPINA1 expression was markedly upregulated in tumor-derived luminal cells (****P < 0.0001), fibroblasts (****P < 0.0001), basal cells (**P < 0.01), and macrophages (P < 0.01) compared to their normal counterparts (Fig. 3D). This result directly validates the bulk RNA-seq finding in Sect. 3.1 that SERPINA1 is specifically overexpressed in BRCA tumor tissues and further clarifies the key cell subsets contributing to this tumor-specific high expression.

Stratifying tumor samples by the median SERPINA1 expression, differential expression analysis (|log₂FC|> 1, FDR < 0.05) identified distinct differentially expressed genes (DEGs) in fibroblasts, basal cells, and macrophages between SERPINA1-high and SERPINA1-low groups (Fig. 3E). KEGG pathway enrichment analysis (adjusted P < 0.05) showed that DEGs in macrophages were enriched in immune-related pathways (e.g., TNF signaling pathway, NF-κB signaling pathway), DEGs in fibroblasts were enriched in ECM-receptor interaction and PI3K-Akt signaling pathway, and DEGs in basal cells were enriched in cell adhesion molecules and T cell receptor signaling pathway (Fig. 3F). These results indicate that SERPINA1 may regulate the biological functions of key cell subsets in the BRCA microenvironment through these pathways, thereby participating in the remodeling of the tumor immune microenvironment.

Association between high SERPINA1 expression and the TIME in BRCA

To further elucidate the immunological differences between high and low SERPINA1 expression groups, we first analyzed the correlation between SERPINA1 expression and a range of immune regulatory factors. The results demonstrated that SERPINA1 expression was positively correlated with a wide array of immune components, including multiple chemokines (e.g., CCL3, CXCL9, CXCL10), immunostimulatory molecules (e.g., CD86, TNFRSF9, ICOS), MHC-related molecules (e.g., TAP1, TAP2, B2M), and immune receptors (e.g., CCR5, CXCR3, CCR2) (Fig. 4A). These findings suggest that SERPINA1 may play a key role in shaping the immune landscape of the BRCA tumor microenvironment.

Fig. 4.

Relationship between high SERPINA1 expression and the TIME in BRCA. A Differential expression analysis of immune-related genes in high and low SERPINA1 expression groups. B Analysis of activity differences in various steps of the cancer-immunity cycle in high and low SERPINA1 expression groups. (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) C Immune cell infiltration levels calculated using six independent algorithms. D Differential expression analysis of immune genes related to CD8 + T cells, dendritic cells, macrophages, NK cells, and Th1 cells. E Differential expression analysis of immune checkpoint genes in high and low SERPINA1 expression groups. (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) F Correlation between SERPINA1 mRNA expression levels and T cell–inflamed score. G Differential expression analysis of TIS-related genes between high and low SERPINA1 expression groups. (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001)

Given the chemokine system's complex and multifaceted nature, correlations between SERPINA1 and individual chemokines alone are insufficient to fully elucidate its immunological effects within the TIME. Therefore, we further evaluated the activity of each step in the cancer-immunity cycle, which offers an integrative view of chemokine-mediated and broader immunoregulatory processes [23, 29].

The analysis revealed that the high SERPINA1 expression group exhibited enhanced activity across multiple critical steps, with particularly significant differences observed in step 3 (priming and activation of immune cells) and step 4 (immune cell recruitment) (Fig. 4B). These results indicate that SERPINA1 may enhance antitumor immunity in BRCA by promoting a more active immune cycle.

Interestingly, although immune activity in most steps of the cancer-immunity cycle (Steps 1–6) was significantly enhanced in the high SERPINA1 expression group, activity in the final step—tumor cell killing (Step 7)—was slightly lower compared to the low expression group. This observation suggests that while SERPINA1 broadly promotes immune activation, other immunosuppressive mechanisms may still constrain the ultimate efficiency of tumor cell elimination.

To investigate the impact of SERPINA1 expression on the distribution of TIICs, we integrated six independent computational algorithms-TIMER, CIBERSORT, quanTIseq, MCP-counter, TISIDB, and TIP-for a comprehensive analysis (Fig. 4C). The results showed that SERPINA1 expression was strongly positively correlated with the infiltration of CD8⁺ T cells, macrophages, and dendritic cells. In contrast, correlations with NK and Th1 cells were relatively weaker.

Building on the analysis of immune cell infiltration, we further assessed the functional status of these immune cells. The results revealed that key immune effector genes associated with CD8⁺ T cells, macrophages, dendritic cells, NK cells, and Th1 cells, such as PRF1, GZMB, CXCL10, CD8A, IFNG, and TBX21, were significantly upregulated in the high SERPINA1 expression group (Fig. 4D). These findings suggest that SERPINA1 may influence immune cell infiltration and enhance the effector functions of these cells.

In addition, significant differences were observed in the expression levels of immune checkpoint molecules between the high and low SERPINA1 expression groups. Specifically, key immune checkpoint genes such as CD274 (PD-L1), PDCD1 (PD-1), CTLA4, LAG3, and TIGIT were markedly upregulated in the high SERPINA1 expression group, further indicating an association with an active immunoregulatory state (Fig. 4E). To quantify the overall trend of tumor immune activity, we assessed the correlation between SERPINA1 expression and the T cell–inflamed score (Fig. 4F). The results revealed a significant positive correlation, suggesting that higher SERPINA1 expression is closely associated with a more immunologically active tumor microenvironment in BRCA. Based on this, we further compared the expression of immune activation–related genes between high and low SERPINA1 expression groups (Fig. 3G). We found that several key T cell–related functional molecules—such as CCL5, CD27, CD274, CXCL9, HLA family genes, PDCD1LG2, and STAT1—were significantly upregulated in the high SERPINA1 expression group.

These findings further support that SERPINA1 enhances immune activity in the tumor microenvironment through multiple T cell–associated pathways, highlighting its pivotal role in immune regulation in BRCA.

Identification of immune-related genes and pathway enrichment analysis between high and low SERPINA1 expression groups

To further explore the relationship between SERPINA1 and the TIME, we calculated the ESTIMATE Score, Immune Score, and Stromal Score for the high and low SERPINA1 expression groups (Fig. 5A). The results showed that all three scores were significantly higher in the high SERPINA1 expression group, indicating that elevated SERPINA1 expression is associated with enhanced immune activity and stromal responses within the BRCA tumor microenvironment.

Fig. 5.

Correlation analysis between SERPINA1 expression levels and TIME characteristics in BRCA. A Differential analysis of immune, stromal, and ESTIMATE scores in high and low SERPINA1 expression groups. B Venn diagram. C KEGG pathway enrichment analysis. D GO pathway enrichment analysis

Conversely, the low SERPINA1 expression group exhibited higher tumor purity (Supplementary Fig. 2A). These findings suggest that SERPINA1 expression is closely linked to the immune contexture of BRCA and may influence the tumor immune ecosystem by modulating the infiltration of immune and stromal cells. To investigate the relationship between SERPINA1 expression and immune/stromal responses further, we stratified samples based on Immune Score, Stromal Score, and SERPINA1 expression level and performed differential gene expression analyses accordingly (Supplementary Fig. 2B). Following differential expression analysis across these groupings, 803 DEGs were identified (Fig. 5B and S1 Table).

Next, we performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses on 803 DEGs. The results revealed that these genes were significantly enriched in multiple immune-related pathways. KEGG analysis showed enrichment in key immune signaling pathways, including the T cell receptor signaling pathway, B cell receptor signaling pathway, and Th17 cell differentiation, suggesting that high SERPINA1 expression may promote antitumor immune responses in BRCA by enhancing these immune pathways (Fig. 5C). In GO enrichment analysis, the significantly enriched biological process (BP) terms were primarily associated with the immune system, immune response, and immune effector processes. This indicates that SERPINA1 upregulation is closely linked to enhanced immune activation, particularly in regulating immune cell activation, proliferation, and effector functions. In addition, cellular component (CC) and molecular function (MF) terms were significantly enriched in immune-related functions, such as the T cell receptor complex and antigen binding, further supporting the role of SERPINA1 in regulating T cell and B cell–mediated immune responses (Fig. 5D).

Construction and validation of a TIME-related prognostic model based on SERPINA1

To evaluate the impact of SERPINA1 and TIME on clinical prognosis in BRCA, we selected 937 patients with complete survival data from the TCGA cohort. We performed univariate Cox regression analysis on the 803 DEGs. This analysis identified 453 genes significantly associated with BRCA prognosis (S2 Table).

Subsequently, we applied LASSO regression to refine these 453 genes further and determine the optimal penalty parameter (λ = 0.0073). At this λ value, the model's feature coefficients stabilized, and seven key genes closely associated with prognosis were selected to construct the prognostic risk model (Fig. 6A&Supplementary Fig. 2C). The Risk Score was calculated using the following formula:

Fig. 6.

Prognostic model of BRCA based on SERPINA1-related immune genes and validation analysis. A Lasso regression. B Distribution of risk scores and survival status in the TCGA cohort. C Survival and ROC curve based on risk groups in the TCGA cohort. D Validation using GSE20685. E Validation using GSE42568

Risk Score = (− 0.2409 × RAC2) + (− 0.2333 × APOBEC3D) + (− 0.2184 × FAM159A) + (− 0.1078 × TFPI2) + (− 0.0892 × ADRB1) + (− 0.0814 × IGHD) + (− 0.0721 × PIGR).

To further assess the association between the Risk Score and BRCA patient prognosis, we stratified patients into high- and low-risk groups based on the median Risk Score. The risk distribution plot demonstrated that a higher Risk Score was closely associated with poorer OS outcomes (Fig. 6B). Kaplan–Meier survival analysis further confirmed that patients in the low-risk Score group had significantly longer OS than those in the high-risk score group. In addition, ROC curve analysis demonstrated that the model had good prognostic predictive performance, with area under the curve (AUC) values of 0.66, 0.68, and 0.65 at 1, 3, and 5 years, respectively (Fig. 6C). These results indicate that the seven-gene risk model can effectively stratify BRCA patients into different survival risk subgroups, and its predictive accuracy is consistent across short-term and long-term follow-up.

To further validate the accuracy of the Risk Score model, we performed external validation using two independent datasets: GSE20685 and GSE42568. In the GSE20685 dataset, Kaplan–Meier survival analysis revealed that patients in the low-risk Score group had significantly better survival probabilities than those in the high-risk score group. ROC curve analysis showed that the AUC values for 1-, 3-, and 5-year survival were 0.63, 0.62, and 0.66, respectively (Fig. 6D). Similarly, in the GSE42568 dataset, Kaplan–Meier analysis demonstrated significantly higher survival probability in the low-risk Score group, and ROC curve analysis yielded AUC values of 0.83, 0.61, and 0.62 for 1-, 3-, and 5-year survival, respectively. These findings indicate that the model has strong predictive ability for short-term survival and maintains stable performance in longer-term prognosis (Fig. 6E), with consistent stratification efficiency across different cohorts (Supplementary Table S5), supporting its reliability as a prognostic tool.

Construction and evaluation of the nomogram

By integrating the Risk Score with clinicopathological variables—including patient age, tumor size (T), lymph node status (N), and disease stage—we constructed a prognostic nomogram for BRCA (Fig. 7A, Supplementary Fig. 2D). Multivariate analysis identified Risk Score, age, and lymph node status as independent prognostic factors. Calibration curve analysis demonstrated a high degree of concordance between the predicted and observed survival probabilities at 1, 3, and 5 years, confirming the model's predictive accuracy (Fig. 7B).

Fig. 7.

Nomogram analysis for BRCA prognosis constructed based on risk model and clinical features. A Nomogram plot. B Calibration curve. C ROC curve

ROC curve analysis further indicated that the nomogram exhibited strong prognostic performance, with AUC values of 0.770, 0.763, and 0.776 for 1-, 3-, and 5-year survival, respectively (Fig. 7C).

Immune microenvironment validation based on the risk model

Previous analyses (Fig. 4) indicated that high SERPINA1 expression is closely associated with an active immune microenvironment. To further validate whether the risk model based on SERPINA1-associated immune genes reflects this immunological profile, we compared the TIME in depth between the high- and low-risk groups.

First, in the analysis of chemokines, immunostimulatory molecules, MHC molecules, and immune receptors, we found that patients in the low-risk group exhibited significantly higher expression levels of related genes, including CCL5, CXCL9, CD86, ICOS, TAP1, TAP2, and B2M (Fig. 8A). These findings are consistent with the immune-activated phenotype observed in the high SERPINA1 expression group.

Fig. 8.

Immune microenvironment differences in BRCA based on the risk model. A Differential expression analysis of immune-related genes in high and low-risk groups. B Analysis of activity differences in various steps of the cancer-immunity cycle in high and low-risk groups. (**p < 0.01, ***p < 0.001) C Immune cell infiltration levels calculated using six independent algorithms. D Differential expression analysis of immune genes related to CD8 + T cells, dendritic cells, macrophages, NK cells, and Th1 cells. (E) Differential expression analysis of immune checkpoint genes in high and low-risk groups. (***p < 0.001) F Differential expression analysis of T cell functional activity-related genes in high and low-risk groups. (G) Differential expression analysis of TIS-related genes in high and low-risk groups. (*p < 0.05, **p < 0.01, ***p < 0.001)

Further analysis of the cancer-immunity cycle revealed that the low-risk group demonstrated more significant activity in key immune processes, particularly in Step 3 (priming and activation of immune cells) and Step 4 (immune cell recruitment) (Fig. 8B, C). In the analysis of immune cell effector functions, the low-risk group exhibited significantly higher expression of key functional genes associated with CD8⁺ T cells, dendritic cells, macrophages, NK cells, and Th1 cells—such as PRF1, GZMB, CD8A, IFNG, and CXCL10 (Fig. 8D). These results suggest the low-risk group possesses more vigorous anti-improve immune effector activity.

Further analysis of immune checkpoint molecules (Fig. 8E) revealed that key inhibitory checkpoint genes, including PD-1, CTLA4, TIGIT, and LAG3, were significantly upregulated in the low-risk group. This suggests that the low-risk group may exist in a state of heightened immune activation while also being subject to regulation by immune inhibitory pathways—providing a rationale for potential immunotherapeutic strategies. In addition, the expression of molecules associated with T cell functional activity, such as CD27, CD8A, CXCL9, and STAT1, was also significantly increased in the low-risk group (Fig. 8F, G), further supporting the notion that this group possesses a more active immune microenvironment and enhanced immune responsiveness.

Drug sensitivity analysis based on the risk model

To evaluate differences in drug responsiveness between Risk Score subgroups, we performed drug sensitivity analysis using “OncoPredict “ package. The results revealed that patients in the low-risk group were significantly more sensitive to standard chemotherapeutic agents, including docetaxel, epirubicin, gemcitabine, and paclitaxel (Fig. 9A).

Fig. 9.

Drug sensitivity analysis of high and low-risk groups. A Sensitivity analysis of conventional chemotherapeutic agents in BRCA for high and low-risk groups. (****p < 0.0001) B Sensitivity analysis of emerging drugs in clinical research for high and low-risk groups. (****p < 0.0001) C Sensitivity analysis of FDA-approved drugs primarily used for other cancer types in high and low-risk groups. (****p < 0.0001)

Further analysis of investigational agents showed that the low-risk group exhibited higher sensitivity to a variety of emerging targeted therapies, including mTOR inhibitors (AZD8055, dactolisib), CDK inhibitors (dinaciclib), HSP90 inhibitors (luminespib), BCL-2 family inhibitors (sabutoclax), and survivin inhibitors (sepantronium bromide) (Fig. 9B).

In addition, when analyzing approved drugs that are primarily used for other cancer types—such as bortezomib, dactinomycin, rapamycin, vinblastine, and vincristine—patients in the low-risk group also exhibited greater sensitivity (Fig. 9C).

We also identified several drugs, including Camptothecin, CDK9 inhibitors (5038, 5576), Eg5 inhibitors (9814), MG.132, Podophyllotoxin bromide, and Staurosporine, which have not yet been approved for clinical use and have not been specifically studied for breast cancer treatment. However, these drugs have shown potential in other types of cancer and may serve as candidate drugs for future research on their impact on SERPINA1 in breast cancer (Supplementary Fig. 3).

Discussion

This study systematically explored the role of SERPINA1 in BRCA and yielded three core findings: first, SERPINA1 is significantly overexpressed in BRCA tissues and correlates with favorable overall survival; second, high SERPINA1 expression shapes an immunologically active tumor microenvironment (TIME), characterized by enhanced infiltration of CD8⁺ T cells, macrophages, and dendritic cells, activation of T/B cell signaling pathways, and upregulation of immune checkpoint molecules (PD-L1, PD-1, CTLA4), a pattern consistent with clinical observations from the KEYNOTE-086 trial, which reported higher response rates to pembrolizumab in PD-L1–positive BRCA patients [41, 42]; third, the constructed seven-gene prognostic model effectively stratifies survival risk across TCGA and external datasets (GSE20685, GSE42568), with low-risk patients exhibiting stronger immune activity and higher sensitivity to chemotherapy (docetaxel, paclitaxel) and targeted agents (mTOR inhibitors). Notably, our data differ from studies in hepatocellular carcinoma—where SERPINA1 modulates TIME via CXCL9/CXCL10 (without linking to prognosis)—by establishing a direct prognostic role and immune regulatory network of SERPINA1 in BRCA. These findings align with prior evidence that increased immune infiltration improves BRCA prognosis [43] and that PD-L1 upregulation predicts response to anti-PD-1 therapy [41], validating the biological rationality of our results. Importantly, the positive correlation between SERPINA1 overexpression and favorable prognosis is not contradictory: SERPINA1 does not directly promote tumor cell proliferation or invasion but rather exerts its antitumor effect indirectly by remodeling the TIME toward an immune-active state, enhancing the recruitment and effector function of antitumor immune cells, which ultimately restrains tumor progression and improves patient survival. This is consistent with the well-established paradigm that immune activation within the TIME is a key determinant of favorable outcomes in BRCA [43, 44].

This study uncovers the tissue-specific immunoregulatory role of SERPINA1 in BRCA, which contrasts sharply with its immunosuppressive functions in other cancer types. In pancreatic cancer, SERPINA1 activates cancer-associated fibroblasts via the IL-6/STAT3 pathway to remodel an immunosuppressive microenvironment [8], while in colorectal cancer, it increases regulatory T cell infiltration and suppresses cytotoxic T lymphocyte function [9]. Notably, our data differ from studies in hepatocellular carcinoma, where SERPINA1 modulates TIME via CXCL9 and CXCL10 [45, 46], by establishing a direct prognostic role and immune regulatory network of SERPINA1 in BRCA. In contrast, our results demonstrate that SERPINA1 promotes an immune-active TIME in BRCA by enhancing antitumor immune cell recruitment and pathway activation. This divergence may stem from intrinsic differences in TIME characteristics: BRCA exhibits higher levels of pro-inflammatory cytokines (IL-2, IFN-γ) and a greater proportion of dendritic cells [5, 43] compared to pancreatic or colorectal cancer, which may rewire the downstream signaling of SERPINA1 from immunosuppression to immunostimulation. This tissue-specificity highlights the necessity of cancer-type-focused investigations and precludes direct extrapolation of SERPINA1-related therapeutic strategies across malignancies.

High SERPINA1 expression was accompanied by notable upregulation of multiple immune checkpoint molecules, including PD-1, PD-L1, and CTLA-4. A similar immunoregulatory pattern has been observed in melanoma, where increased PD-L1 expression is closely associated with response to immunotherapy [47, 48]. These findings suggest that the immunological features associated with SERPINA1 expression in BRCA may point to its potential utility as a predictive biomarker for immunotherapy response.

Interestingly, high SERPINA1 expression is associated with both enhanced immune activation and constrained tumor cell killing [23, 44]. This paradox is resolved by the "adaptive immune resistance" mechanism [49, 50]: SERPINA1 first drives robust antitumor immunity, but as the immune response intensifies, tumor cells initiate a negative feedback loop by upregulating immune checkpoints. This feedback restrains the cytotoxic activity of infiltrating immune cells, explaining why high immune activation does not translate to complete tumor elimination. This mechanism also provides a theoretical basis for personalized treatment: low-risk patients (high immune activation and checkpoint upregulation) may benefit from chemotherapy and targeted therapy combined with immune checkpoint inhibitors (ICIs) to block adaptive resistance, while high-risk patients (cold tumor phenotype) require TIME-remodeling combinations to convert "cold" to "hot" tumors [19, 51–54].

Notably, SERPINA1 has also been shown to promote tumor growth and immune evasion through modulation of the PI3K/Akt/mTOR pathway in other cancer types [14, 55]. In this study, the co-occurrence of high SERPINA1 expression with low-risk scores and enhanced immune activity suggests that SERPINA1 may concurrently regulate both the mTOR/PI3K axis and antitumor immunity. These findings provide a novel theoretical basis for exploring combined mTOR inhibition and immunotherapy in BRCA.

This study has several limitations. First, the lack of DCA limits direct evidence for the model’s net clinical benefit in practical decision-making, which will be addressed by integrating real-world clinical data in future studies. Second, the retrospective design and limited clinical samples may introduce selection bias, necessitating validation in large-scale prospective cohorts with diverse populations. Third, mechanistic exploration remains preliminary: while scRNA-seq and bulk RNA-seq analyses indicate associations between SERPINA1 and immune activation, the specific molecular pathways (e.g., direct interaction with immune cell receptors or protease inhibition-mediated chemokine regulation) require verification via in vitro/in vivo functional experiments. Finally, drug sensitivity predictions and immunotherapy response correlations are based on computational analyses, lacking experimental validation through in vitro drug susceptibility assays.

Building on these limitations, future research could further expand the clinical translation value of our findings. Specifically, preclinical models could be used to assess whether SERPINA1 expression levels predict responses to ICIs [56, 57] and to explore its potential as a predictive biomarker for chemosensitization to develop SERPINA1-based combination treatment strategies. Furthermore, the prognostic model developed in this study could be improved by integrating multi-omics data, including DNA methylation, non-coding RNA profiles, and spatial transcriptomics, to further enhance the accuracy of patient outcome predictions and address the current gap in understanding the spatial distribution of SERPINA1-expressing cells within the tumor microenvironment.

Conclusion

This study systematically highlights the vital role of SERPINA1 in immune regulation and prognosis in BRCA. The prognostic model built on SERPINA1-related immune genes provides a new scientific basis for patient risk stratification and clinical decision-making. Our findings establish the theoretical foundation for developing more effective personalized immunotherapy strategies and combination treatments, offering new hope for improved prognosis and survival outcomes in BRCA patients.

Author contribution

This statement replaces any existing authorship contribution statements in the manuscript and will be the official version for publication. X.X. (Xiaolin Xia) was responsible for collecting paired breast cancer and adjacent normal tissue samples from clinical patients, performing experimental validations (including Western blot and qPCR assays), and collating raw experimental data. S.C. (Shuai Chen) designed the overall study framework, conducted key bioinformatics analyses (including TCGA/GEO dataset processing, GSVA pathway enrichment analysis, LASSO regression for prognostic model construction, and immune cell infiltration analysis using TIMER/CIBERSORT algorithms), and wrote the main text of the manuscript. Z.C. (Zhou Chen) supervised the entire research process, verified the integrity and accuracy of experimental and bioinformatics data, revised the manuscript for academic rigor, and coordinated communications with the institutional ethics committee. All authors (X.X., S.C., Z.C.) reviewed the manuscript draft, confirmed the validity of results, approved the final version for submission, and assume accountability for the integrity of all research content.

Funding

This study was supported by the Yiyang Science and Technology Innovation Project (No. [2022]08), the Natural Science Foundation of Hunan Province (Higher Education, Joint) (No. 2025JJ90083), and the Natural Science Foundation of Hunan Province (Regional Joint) (No. 2023JJ503602).

Data availability

All relevant data are within the manuscript and its Supporting Information files.

Declarations

Informed consent

Informed consent was obtained from all individual participants included in the study.

Publication consent

We confirm that all authors have read and approved the final version of the manuscript and agree to its submission to this journal. We also declare that the work described is original and has not been published elsewhere, nor is it under consideration for publication elsewhere. All authors have made substantial contributions to the conception, design, execution, or interpretation of the work, and have approved the submitted version. We acknowledge that any unauthorized use of copyrighted material within this manuscript is the responsibility of the authors.

Competing interests

The authors declare no competing interests.