Abstract

Background

   Triple-negative breast cancer (TNBC) is a heterogeneous disease with a
   worse prognosis. Despite ongoing efforts, existing therapeutic
   approaches show limited success in improving early recurrence and
   survival outcomes for TNBC patients. Therefore, there is an urgent need
   to discover novel and targeted therapeutic strategies, particularly
   those focusing on the immune infiltrate in TNBC, to enhance diagnosis
   and prognosis for affected individuals.

Methods

   The gene co-expression network and gene ontology analyses were used to
   identify the differential modules and their functions based on the GEO
   dataset of [34]GSE76275. The Weighted Gene Co-Expression Network
   Analysis (WGCNA) was used to describe the correlation patterns among
   genes across multiple samples. Subsequently, we identified key genes in
   TNBC by assessing genes with an absolute correlation coefficient
   greater than 0.80 within the eigengene of the enriched module that were
   significantly associated with breast cancer subtypes. The diagnostic
   potential of these key genes was evaluated using receiver operating
   characteristic (ROC) curve analysis with three-fold cross-validation.
   Furthermore, to gain insights into the prognostic implications of these
   key genes, we performed relapse-free survival (RFS) analysis using the
   Kaplan-Meier plotter online tool. CIBERSORT analysis was used to
   characterize the composition of immune cells within complex tissues
   based on gene expression data, typically derived from bulk RNA
   sequencing or microarray datasets. Therefore, we explored the immune
   microenvironment differences between TNBC and non-TNBC by leveraging
   the CIBERSORT algorithm. This enabled us to estimate the immune cell
   compositions in the breast cancer tissue of the two subtypes. Lastly,
   we identified key transcription factors involved in macrophage
   infiltration and polarization in breast cancer using transcription
   factor enrichment analysis integrated with orthogonal omics.

Results

   The gene co-expression network and gene ontology analyses revealed 19
   modules identified using the dataset [35]GSE76275. Of these, modules 5,
   11, and 12 showed significant differences between in breast cancer
   tissue between TNBC and non-TNBC. Notably, module 11 showed significant
   enrichment in the WNT signaling pathway, while module 12 demonstrated
   enrichment in lipid/fatty acid metabolism pathways. Subsequently, we
   identified SHC4/KCNK5 and ABCC11/ABCA12 as key genes in module 11 and
   module 12, respectively. These key genes proved to be crucial in
   accurately distinguishing between TNBC and non-TNBC, as evidenced by
   the promising average AUC value of 0.963 obtained from the logistic
   regression model based on their combinations. Furthermore, we found
   compelling evidence indicating the prognostic significance of three key
   genes, KCNK5, ABCC11, and ABCA12, in TNBC. Finally, we also identified
   the immune cell compositions in breast cancer tissue between TNBC and
   non-TNBC. Our findings revealed a notable increase in M0 and M1
   macrophages in TNBC compared to non-TNBC, while M2 macrophages
   exhibited a significant reduction in TNBC. Particularly intriguing
   discovery emerged with respect to the transcription factor FOXM1, which
   demonstrated a significant regulatory role in genes positively
   correlated with the proportions of M0 and M1 macrophages, while
   displaying a negative correlation with the proportion of M2 macrophages
   in breast cancer tissue.

Conclusion

   Our research provides new insight into the biomarkers and immune
   infiltration of TNBC, which could be useful for clinical diagnosis of
   TNBC.

   Keywords: gene co-expression network, immune infiltration, biomarkers,
   macrophages, triple negative breast cancer

Introduction

   Breast cancer (BC) is the most prevalent malignancy affecting women
   globally and remains a significant contributor to cancer-related
   deaths. Over 3 million new cases of breast cancer and 1 million deaths
   will occur each year worldwide by the year 2040 ([36]Arnold et al.,
   2022). TNBC constitutes a particularly aggressive and invasive subtype
   characterized by the absence of estrogen receptor (ER), progesterone
   receptor (PR), and human epidermal growth factor receptor 2 (HER2)
   expression ([37]Agarwal et al., 2016). TNBC occurs in about 15%–20% of
   all breast cancer, which are different from non-TNBC with prognosis and
   therapeutic targets. Compared with other breast cancer subtypes, TNBC
   is always related to worse prognosis and lower overall survival rate
   ([38]Dietze et al., 2015). The lack of hormone receptors and HER2
   expression can make it more difficult to cure so that making
   therapeutic targeting difficult and combination therapy is needed
   ([39]Wang Y. et al., 2019; [40]Dong et al., 2018). The current main
   treatment of TNBC is still chemotherapy, but a significant number of
   patients are drug resistant, leading to poor therapeutic effect, which
   was not satisfactory ([41]Hu et al., 2023; [42]Yang et al., 2021).
   Consequently, the identification of novel therapeutic targets holds
   paramount importance in augmenting the prognosis and treatment efficacy
   of TNBC ([43]Zhu et al., 2023).

   Weighted gene co-expression network analysis (WGCNA) has demonstrated
   wide-ranging applications in biomarker discovery, including, but not
   limited to, laryngeal cancer ([44]Liu et al., 2019), lung cancer
   ([45]Wang S. et al., 2023) and advanced gastric cancer ([46]Wang W-J.
   et al., 2019). Previous studies have revealed that AMD1, EN1, and VGLL1
   are likely to contribute to breast cancer progression and an
   unfavorable prognosis ([47]Yang et al., 2021). Additionally, IL6ST,
   HMGA1, FOXM1, and MYBL2 have been identified as potentially playing an
   important role in TNBC progression using the TNBC gene expression
   dataset [48]GSE76275 ([49]Jia et al., 2021; [50]Fiscon et al., 2021).
   Despite this progress, the potential biomarkers and their mechanisms
   involved in breast cancer subtypes have not been fully elucidated.
   Furthermore, the involvement of these potential marker genes in
   tumor-associated macrophages (TAMs) has not been reported. TAMs, one of
   the main cell types in the tumor immune microenvironment, play a
   pivotal role in cancer progression ([51]Wang XQ. et al., 2023; [52]Chen
   et al., 2023). TAMs are innate immune effector cells that are recruited
   to tumor tissues, contributing to tumor growth and metastasis by
   promoting angiogenesis and suppressing adaptive immunity ([53]Niu et
   al., 2016). TAMs represent a heterogeneous and plastic population,
   within which polarized TAMs can be identified as M1-and M2-like
   macrophages ([54]Gordon and Taylor, 2005). Recently, some clinical and
   experimental research has discovered that M1-like macrophages release
   proinflammatory cytokines and chemokines, such as tumor necrosis factor
   (TNF-α), interleukin (IL-1β), and CXCL10, exerting antitumor activity
   ([55]Pe et al., 2022). In contrast, M2-like macrophages produce a high
   number of anti-inflammatory factors, which may lead to the immune
   escape of tumor cells and contribute to breast cancer progression
   ([56]Zhang et al., 2024). The immunosuppressive role of M2 macrophages
   in TNBC indicates their potential as therapeutic biomarkers ([57]Zhang
   et al., 2024). However, differences between M1-and M2-like macrophages
   in TNBC and non-TNBC, especially the biomarkers of M1-and M2-like
   macrophages in TNBC, are rarely reported. Consequently, a comprehensive
   characterization of the immune infiltrate in TNBC holds significant
   promise for identifying patients most likely to benefit from
   immunotherapy and identifying resistance factors that could serve as
   potential therapeutic targets.

   In this study, we conducted a comprehensive investigation aimed at
   elucidating the molecular characteristics and immune microenvironment
   of breast cancer subtypes, with a particular focus on TNBC and
   non-TNBC. To begin, we employed the gene co-expression network and gene
   ontology analyses to identify the differential modules and the function
   of these modules based on the GEO dataset of [58]GSE76275. Next, we
   identified key genes associated with these subtypes and evaluated their
   potential as diagnostic biomarkers through ROC curve analysis.
   Furthermore, we performed relapse-free survival (RFS) analysis on these
   key genes using the Kaplan-Meier plotter online tool to determine their
   prognostic value and clinical relevance. Additionally, we estimated the
   immune cell compositions in breast cancer tissue between TNBC and
   non-TNBC using the CIBERSORT algorithm. Finally, we identified key
   transcription factors involved in macrophage infiltration and
   polarization in breast cancer through transcription factor enrichment
   analysis by integrating orthogonal omics data.

Materials and methods

Data collection and processing

   The [59]GSE76275 dataset was retrieved from the GEO database and
   comprises 198 triple-negative breast cancer tissue samples and 67
   non-triple-negative breast cancer tissue samples ([60]Burstein et al.,
   2015). To ensure complete and high-quality data for subsequent
   analyses, a filtering procedure was carefully conducted to exclude
   entries lacking age and BMI information. This rigorous filtering step
   ensured that only samples with the necessary clinical attributes were
   retained for further investigation, thereby enhancing the reliability
   and robustness of our analysis.

Weighted gene co-expression network analysis (WGCNA)

   Weighted gene co-expression network Analysis (WGCNA) for microarray
   data was constructed and analyzed using the WGCNA package in R
   ([61]Zhang and Horvath, 2005). In this study, we utilized 20,529 genes
   to construct the gene co-expression network. Confounding factors,
   including age and BMI, were adjusted for in the construction of the
   gene co-expression network. A power threshold of 8 was selected to
   calculate the weighted adjacency matrix, with the thresholding
   parameter defined using a scale-free topology with a cutoff R^2 = 0.8.
   This cutoff was chosen to be the closest to or slightly above 0.8 to
   establish a biologically meaningful and stable network for further
   analysis. We identified gene modules using the ‘hybrid’ method with
   parameters mergeCutHeight = 0.25 and minModuleSize = 40. Modules were
   identified as branches in the dendrogram using the Dynamic Tree Cut
   algorithm ([62]Rezaie et al., 2023). Subsequently, we used topological
   overlap measures (TOM) to reveal the interconnectedness and functional
   relationships among genes within the network ([63]Li and Horvath,
   2007).

Evaluation of immune cells between TNBC and Non-TNBC

   To identify the different immune cell compositions of breast cancer
   tissue between TNBC and non-TNBC, we employed the CIBERSORT algorithm,
   using the leukocyte signature matrix (LM22) as the reference gene
   expression signatures ([64]Newman et al., 2015). CIBERSORT is an
   algorithm that utilizes a reference gene expression signature matrix
   (LM22), representing 22 immune cell types, to deconvolve complex tissue
   expression data into estimated immune cell fractions ([65]Newman et
   al., 2015). It employs support vector regression to minimize errors in
   predicting cell type fractions, even when cell types have highly
   similar gene expression profiles. We obtained bulk-RNA sequencing data
   from breast cancer tissue for TNBC and non-TNBC from the [66]GSE76275
   dataset for our analysis. Using R software, we fully analyzed and
   visualized the abundance and proportion of immune cell members among
   different groups.

GO and KEGG enrichment analyses

   The Gene Ontology (GO) and Kyoto Encyclopedia of Genes and genomes
   (KEGG) pathway enrichment analysis were performed using
   “clusterProfile” package in R software ([67]Yu et al., 2012). The
   Benjamin-Hochberg approach was used to correct multiple tests and
   select the significant terms and pathways. The adjusted p-value <0.05
   was used as a threshold of significance for the enriched terms and
   pathways for target genes.

Transcription factor enrichment analysis by orthogonal omics integration

   To obtain the key transcription factors of genes that associated with
   macrophage infiltration and polarization in breast cancer, we performed
   transcription factor enrichment analysis by orthogonal omics
   integration following Chip-X Enrichment Analysis Version 3 (ChEA3)
   ([68]Keenan et al., 2019). TF enrichment analysis (TFEA) prioritizes
   transcription factors based on the overlap between given lists of
   differentially expressed genes, and previously annotated TF targets
   assembled from multiple resources. The multiple resources including
   ChIP-seq experiments from ENCODE, ReMap, and individual publications;
   co-expression of TFs with other genes based on processed RNA-seq from
   GTEx and ARCHS4; co-occurrence of TFs with other genes by examining
   thousands of gene lists submitted to the tool [69]Enrichr; and gene
   signatures resulting from single TF perturbations followed by
   genome-wide gene expression experiments.

Prognostic analysis of triple negative breast cancer

   A Kaplan Meier plotter ([70]https://kmplot.com) database, discovery
   based on meta-analysis, and validation of online survival biomarker
   tools were used to study the survival of TNBC in GEO, the European
   Archives of Genomic Phenomena (EGA), TCGA, and METRIC databases.

Protein-protein interaction network analysis

   The search tool for retrieving interacting genes is a database of known
   and predicted protein-protein interactions that can be used to predict
   and track the protein–protein interactions network. This study used the
   STRING database to construct the PPI network of macrophage infiltration
   and polarization related genes in breast cancer.

Statistical analysis

   Data statistical analysis and visualization were performed using R
   software. The correlation among different continuous variables were
   obtained by Spearman’s correlation coefficient. Multiple comparisons
   between categorical variables using ANOVA analysis, and t.test were
   applied for statistical analysis between different groups. The p-value
   <0.05 was considered as significantly statistical difference in this
   study.

Results

Preprocessing of data and descriptive statistics analysis

   To obtain more complete clinical information on TNBC and non-TNBC, we
   removed samples with incomplete age and BMI data. A total of 201
   samples, comprising 149 TNBC and 52 non-TNBC samples, were retained for
   subsequent analyses. All the samples were recorded age (average: 56.5,
   range: 26-87) and BMI (average: 28.1, range: 16-56) ([71]Figure 1A;
   [72]Supplementary Table S1). To assess any potential disparities in age
   and BMI between TNBC and non-TNBC breast cancer tissues, we conducted a
   comparative analysis, which indicated no significant differences in age
   or BMI distribution between the two subtypes ([73]Figure 1B). To
   control for the effect of age and BMI on gene expression, we used a
   linear mixed model to correct for confounding factors, including age
   and BMI. After correction, age and BMI had minimal effect on gene
   expression, thus preserving the integrity of our downstream gene
   expression analysis ([74]Supplementary Figure S1). Lastly, to examine
   global gene expression patterns between TNBC and non-TNBC, we conducted
   principal component analysis (PCA) on the [75]GSE76275 dataset
   comprising 20,529 genes ([76]Figure 1C). The PCA revealed distinct
   clustering of TNBC and non-TNBC breast cancer tissue, underscoring
   substantial differences in gene expression profiles between these two
   breast cancer subtypes.

FIGURE 1.

   [77]FIGURE 1
   [78]Open in a new tab

   The detail information of breast cancer tissue between TNBC and
   non-TNBC. (A) The distribution of age and BMI in TNBC and non-TNBC. (B)
   Boxplot of age and BMI between TNBC and non-TNBC. The “ns” represents
   there is no significant difference between Non-TNBC and TNBC group. (C)
   Scatter plot of the first two principal component vectors of the gene
   expression profiles of samples from TNBC and non-TNBC, which are
   highlighted using different colors. Red represents TNBC, blue
   represents non-TNBC.

Identification of the TNBC related gene co-expression modules

   To elucidate the potential functions and underlying mechanisms of key
   genes in breast cancer tissue between TNBC and non-TNBC, we constructed
   a weighted gene co-expression network using the [79]GSE76275 dataset.
   For creating a robust co-expression network, we applied a R^2 cutoff of
   0.80 and set the soft-threshold power (β) to 8. Consequently, we
   constructed adjacency matrices to capture comprehensive co-expression
   information across the network. This analysis successfully derived 19
   gene modules, each containing more than 40 genes ([80]Figures 2A, B).
   Subsequently, we built additional co-expression networks associated
   with tumor grade, tumor size, and breast cancer subtypes by employing
   the Pearson correlation coefficient within these 19 gene modules
   through WGCNA analysis. Our results revealed that modules 5, 11, and 12
   displayed significant correlations with the breast cancer subtypes, as
   evidenced by correlation coefficients greater than 0.5 and adjusted
   p-values less than 0.05 ([81]Supplementary Table S2). Notably, module
   11 also exhibited a significant negative correlation with tumor grade.
   However, no significant correlations were observed between the gene
   modules and tumor size ([82]Figure 2C). These findings indicate the
   potential functional relevance of the identified gene modules in breast
   cancer pathogenesis and further underscore the importance of these
   genes in distinguishing between TNBC and non-TNBC subtypes.

FIGURE 2.

   [83]FIGURE 2
   [84]Open in a new tab

   Weighted gene co-expression network constructed in breast cancer tissue
   between TNBC and non-TNBC. (A) Gene co-expression network module in
   breast cancer tissue between TNBC and non-TNBC. (B) The numbers of
   modules and genes in weighted gene co-expression network modules based
   on the dataset of [85]GSE76275. (C) A heatmap of the correlations
   between 19 module principal components (PCs) and 3 cancer-related
   traits including tumor grade, tumor size and the subtypes of breast
   cancer.

Function annotation and key genes of TNBC related modules

   Next, we performed bioinformatics analysis of breast cancer-related
   modules using clusterProfiler package in R software. The results
   highlighted two distinct modules, module 11 and module 12, each
   associated with specific biological processes ([86]Figure 3A;
   [87]Supplementary Table S3). Genes within module 11 were found to be
   involved in the regulation of canonical Wnt signaling, tube formation,
   and epithelium development, while those in module 12 were primarily
   associated with fatty acid metabolic processes, small molecule
   catabolic processes, and organic acid catabolic processes. ([88]Figure
   3B; [89]Supplementary Table S3). These results indicated that these
   pathways are essential to involve in the progression of TNBC. Moreover,
   we identified 9 genes within the enriched modules by selecting those
   with module membership (MM) absolute correlation coefficients greater
   than 0.80, and significant correlations with breast cancer subtypes in
   module 11. The overlapping genes between these 9 genes and the top 10
   genes with significant correlation with breast cancer subtypes were
   considered as key genes in module 11. The identification process for
   the 2 key genes in module 12 is consistent with that used for
   identifying key genes in module 11. Specifically, SHC4 and KCNK5 were
   identified as crucial key genes in module 11, while ABCC11 and ABCA12
   were identified as crucial key genes in module 12 ([90]Figures 3C, D).
   Notably, SHC4 and KCNK5 exhibited significantly higher gene expression
   levels in TNBC compared to non-TNBC ([91]Supplementary Figure S2),
   while ABCC11 and ABCA12 displayed lower expression levels in TNBC
   ([92]Supplementary Figure S3). Finally, we evaluated the diagnostic
   potential of these key genes using receiver operating characteristic
   (ROC) curve analysis with three-fold cross-validation. The results
   showed that the AUC of the three verification sets in the logistic
   model constructed by the 3-fold cross-validation method was 0.973,
   0.937 and 0.978 with an average AUC of 0.963 ([93]Figure 3E). Taken
   together, these results suggest that these four key genes play key
   roles in the progression of TNBC which may contribute potential targets
   for the diagnosis, treatment, and prognosis assessment of TNBC.

FIGURE 3.

   [94]FIGURE 3
   [95]Open in a new tab

   Function annotation and key genes of TNBC related modules. (A, B) The
   top 10 enrichment pathways of genes in module11 and module12. (C, D)
   Venn diagrams showing common genes between genes A and genes B in
   module 11 and module 12, respectively. Genes A represents with exhibit
   an absolute correlation coefficient >0.80 within the eigengene of the
   enriched module and significantly correlated with the subtypes of
   breast cancer, and Genes B represents top 10 genes significantly
   associated with the subtypes of breast cancer. (E) The ROC curve of
   four essential gene combinations is based on the logistic regression
   model.

The association between key targets and prognosis of TNBC

   To elucidate the prognostic implications of the four identified genes
   in TNBC, we performed a comprehensive prognostic analysis using the
   Kaplan-Meier plotter online tool. The results indicated that three out
   of the four genes significantly impact TNBC prognosis, as shown in
   [96]Figure 4. Specifically, KCNK5, ABCC11, and ABCA12 emerged as key
   prognostic markers with substantial implications for clinical
   management and treatment decisions in TNBC ([97]Figure 4). These
   findings underscore the potential clinical relevance of these three
   target genes in predicting patient outcomes and highlight their utility
   as prognostic indicators in TNBC.

FIGURE 4.

   [98]FIGURE 4
   [99]Open in a new tab

   The KM plotter online cancer survival analysis tool
   ([100]http://kmplot.com/analysis/) to evaluate the Relapse-free
   survival (RFS) in six key genes. (A) SHC4, (B) KCNK5, (C) ABCC11 and
   (D) ABCA12 in samples with TNBC on Kaplan Meier plotter analysis.

Identification of the immune cell compositions between TNBC and Non-TNBC

   The CIBERSORT deconvolution algorithm was exploited to access the
   immune cell compositions in breast cancer tissue between TNBC and
   non-TNBC. We summarized the results obtained from the remaining 201
   breast cancer tissue between TNBC and non-TNBC in [101]Figure 5A.
   Compared to breast cancer tissue of non-TNBC, the breast cancer tissue
   in TNBC exhibited higher infiltration of M0 and M1 macrophages, lower
   infiltration of M2 macrophages. In addition to macrophages, there were
   seven immune cell types also exhibiting significantly differential
   proportions in breast cancer tissue between TNBC and non-TNBC including
   plasma cells, T cells CD8, T cells CD4 memory activated, NK cells
   resting, mast cells resting, eosinophils and neutrophils ([102]Figure
   5B). Nonetheless, none of these immune cell types exhibited proportions
   greater than those of macrophages, suggesting that the infiltration and
   polarization of tumor-related macrophages are crucial distinguishing
   features in breast cancer tissue between TNBC and non-TNBC.
   Additionally, we calculated the Pearson correlation coefficient between
   different macrophages and found predominantly negative and significant
   correlations in breast cancer tissue between TNBC and non-TNBC
   ([103]Figure 5C). Finally, we depicted the distribution of different
   macrophage types in TNBC and non-TNBC breast cancer tissues
   ([104]Figure 5D). On average, TNBC samples showed higher proportions of
   M0 macrophages (20.5%) and M1 macrophages (12.1%) compared to non-TNBC,
   where M0 and M1 macrophages averaged 15.9% and 9.79%, respectively.
   Conversely, the average proportion of M2 macrophages in TNBC (13.0%)
   was notably lower than in non-TNBC (18.6%). Together, these findings
   highlight macrophage infiltration and polarization as crucial
   differentiating features between TNBC and non-TNBC breast cancer
   tissue.

FIGURE 5.

   [105]FIGURE 5
   [106]Open in a new tab

   The immune cell compositions in breast cancer tissue between TNBC and
   non-TNBC. (A) The fraction of 22 subsets of immune cells in breast
   cancer tissue between TNBC and non-TNBC. (B) The violin graph shows the
   difference of immune infiltration between TNBC and non-TNBCs. Red
   represents TNBC, Blue represents non-TNBC. (C) Pearson correlation
   coefficient between the different macrophages in breast cancer tissue.
   Red represents positive correlation. Blue represents negative
   correlation. (D) The proportion of distribution of different
   macrophages in breast cancer tissue.

Identification of transcription factors associated with immune infiltration

   To identify transcription factors that affect macrophage infiltration
   and polarization in breast cancer tissue, we first identified 179 genes
   displaying a positive correlation with the proportions of M0 and M1
   macrophages, while concurrently exhibiting a negative correlation with
   the proportion of M2 macrophages. Notably, 47 of these genes
   demonstrated significant differential expression between TNBC and
   non-TNBC samples ([107]Figure 6A). Subsequently, a Protein-Protein
   Interaction (PPI) network analysis was conducted to explore the
   interaction patterns among the 47 differentially expressed genes,
   resulting in the identification of 24 genes exhibiting substantial
   connectivity degrees based on the STRING database ([108]Figure 6B).
   These results indicated that these 24 genes were considered potentially
   crucial elements in the network associated with tumor-associated
   macrophages (TAMs) in breast cancer tissues, specifically
   distinguishing between TNBC and non-TNBC. Further investigations
   focused on transcription factor enrichment analysis, which unveiled the
   top 10 transcription factors governing the expression of the 24
   identified genes. Among these, FOXM1 stood out due to its significant
   differential expression in breast cancer tissues between TNBC and
   non-TNBC, being notably upregulated in TNBC samples ([109]Figure 6C;
   [110]Supplementary Figure S4). Finally, we conducted functional
   enrichment analysis on the genes regulated by FOXM1 and found that they
   were significantly enriched in the nuclear division, organelle fission,
   and chromosome segregation pathway ([111]Figure 6E). Notably, a
   substantial positive correlation emerged between the proportion of
   FOXM1 and M1 macrophages, while conversely, a significant negative
   correlation was observed with the proportion of M2 macrophages
   ([112]Figures 6F, G). These results emphasized the crucial role of
   FOXM1 in the progression of TNBC, modulating genes significantly
   associated with macrophage infiltration and polarization.

FIGURE 6.

   [113]FIGURE 6
   [114]Open in a new tab

   The key transcription of macrophage infiltration and polarization in
   breast cancer tissue. (A) The intersection of DEGs and TAMs-related
   genes in breast cancer tissue. DEGs: the differential expression genes
   in breast cancer tissue between TNBC and non-TNBC. (B) PPI analysis of
   the intersection genes of DEGs and TAMs-related genes. (C) The top 10
   transcription factors of 24 genes with high connectivity. (D)
   Visualization of FOXM1 and their corresponding target gene networks.
   (E) The top 10 enrichment pathways of target genes of FOXM1. (F) The
   correlation between FOXM1 and the proportion of M1 macrophages. (G) The
   correlation between FOXM1 and the proportion of M2 macrophages.

Discussion

   The subtypes of breast cancer not only have different clinical
   manifestations, but also have different prognostic significance. Among
   these subtypes, triple-negative breast cancer (TNBC) has garnered
   significant attention in both clinical and experimental research. Due
   to its high-risk biological features and limited specific treatment
   options. Consequently, there is a pressing need to gain a deeper
   understanding of TNBC biology to identify potential therapeutic
   targets. In this study, we aimed to recognize meaningful prognostic
   biomarkers for TNBC by analyzing the [115]GSE76267 dataset using
   bioinformatic methods. Firstly, we constructed weighted gene
   co-expression networks based on gene expression profiles of breast
   cancer tissue, considering 149 TNBC and 52 non-TNBC cases. By
   identifying key genes and modules within these networks, we sought to
   pinpoint specific genes associated with TNBC prognosis. The prognostic
   significance of these key targets was validated using the Kaplan-Meier
   tool, providing valuable insights into their potential clinical
   relevance. Furthermore, we also investigated the immune cell
   composition of breast cancer tissue between TNBC and non-TNBC.
   Additionally, we conducted an in-depth analysis to identify the
   transcription factors responsible for regulating genes associated with
   macrophage infiltration and polarization in breast cancer tissue.

   Utilizing WGCNA, we diligently investigated the associations between
   gene modules and subtypes of TNBC. Our analysis revealed a substantial
   correlation between modules 5, 11, and 12 and the specific TNBC
   subtypes, as evidenced by correlation coefficients exceeding 0.5 and
   adjusted p-values below 0.05. These compelling findings indicate that
   the genes encompassed within these modules play pivotal roles in the
   intricate progression of TNBC. Of particular interest, module 11
   emerged as an enriched hub of biological processes, encompassing the
   regulation of canonical Wnt signaling, tube formation, and epithelium
   development. Notably, each of these processes has been extensively
   implicated in the pathogenesis of TNBC. For instance, Wnt signaling
   regulates a variety of cellular processes, including cell fate,
   differentiation, proliferation, and stem cell pluripotency.
   Perturbations in Wnt signaling have been implicated in the progression
   of TNBC, signifying its functional relevance in this aggressive breast
   cancer subtype ([116]Pohl et al., 2017). In previous studies, it has
   been reported that tube-formation assay disclosed the function of PCAT6
   on angiogenesis ([117]Dong et al., 2020). These results indicated that
   tube formation mediated angiogenesis is involved in the progression of
   TNBC. Concurrently, module 12 emerged as a vital repository of genes
   primarily involved in fatty acid metabolic processes, small molecule
   catabolic processes, and organic acid catabolic processes. The results
   demonstrated that fatty acid metabolic process might be potential
   differential pathways in breast cancer tissue between non-TNBC and
   TNBC. Fatty acid synthesis and fatty acid oxidation are generally
   viewed as counterparts in metabolic reprogramming of tumor cells
   ([118]Ma et al., 2018). In previous reported that inhibition of fatty
   acid oxidation as a therapy for MYC-overexpressing triple-negative
   breast cancer. In summary, our rigorous WGCNA analysis unraveled
   distinct gene modules significantly correlated with specific TNBC
   subtypes, thereby illuminating key biological processes pertinent to
   the intricate landscape of TNBC progression. The identification of
   these critical pathways and their potential therapeutic implications
   may inform the development of targeted treatments and precision
   medicine approaches, thus holding promise for improving clinical
   outcomes in TNBC patients.

   Our analysis pinpointed SHC4 and KCNK5 as essential genes within module
   11, exhibiting significant correlations with the subtypes of breast
   cancer. These findings strongly suggested that SHC4 and KCNK5 play
   crucial roles in the progression of TNBC. The Src homology and collagen
   (SHC) family is one of the most studied adaptor protein families
   consisting of four members, SHC1, SHC2, SHC3, and SHC4 ([119]Ahmed and
   Prigent, 2017). SCH4 could promote tumor proliferation and metastasis
   by activating STAT3 signaling in hepatocellular carcinoma ([120]Zhang
   et al., 2022). While SHC4 has been implicated in promoting tumor
   proliferation and metastasis by activating STAT3 signaling in
   hepatocellular carcinoma, its specific involvement in TNBC remains
   relatively unexplored, prompting our investigation into its potential
   role in regulating tumor proliferation and metastasis in TNBC
   progression. Furthermore, KCNK channels (also known as K2P, for
   two-pore-domain potassium channels) are potassium-selective channels
   that tend to be constitutively open ([121]Goldstein et al., 1996).
   KCNKs may regulate breast cancer progression via modulating immune
   response which can serve as ideal prognostic biomarkers for breast
   cancer ([122]Zou et al., 2022). Our results further supported that
   KCNKs as potential biomarker in triple negative breast cancer. In
   addition to SHC4 and KCNK5, we also identified ABCC11 and ABCA12 genes
   as key players associated with the progression of TNBC. ATP-binding
   cassette (ABC) transporters are membrane proteins that efflux various
   compounds from cells, including chemotherapeutic agents, and are known
   to affect multidrug resistance ([123]Locher, 2016). ABC family may be a
   useful tool in determining personalized TNBC treatment
   ([124]Makuch-Kocka et al., 2023; [125]Park et al., 2006). In our study
   further supported ABC family including ABCC11 and ABCA12 genes plays an
   important role in the progression of TNBC.

   To evaluate the discriminative power of the four-gene combination
   (SHC4, KCNK5, ABCC11, and ABCA12), we utilized the receiver operating
   characteristic (ROC) curve in a logistic model constructed through
   3-fold cross-validation. The resulting average area under the curve
   (AUC) of 0.963 underscores the potential of these genes as effective
   discriminators between TNBC and non-TNBC. Moreover, through prognostic
   analysis using these key markers, we found that KCNK5, ABCC11, and
   ABCA12 significantly impact the prognosis of TNBC. Importantly,
   previous studies have already reported KCNK5 and ABCC11 as prognostic
   signatures in breast cancer. Specifically, high expression levels of
   ABCC11 were associated with worse disease-free survival in patients
   with HER2+ and triple-negative tumor subtypes ([126]Yamada et al.,
   2013). Additionally, KCNK genes have been identified as prognostic
   signatures for breast cancer, including TNBC ([127]Zou et al., 2022).
   ABCA12 is a highly expressed gene in cancer tissues and cells and has
   been identified as a key gene related to the prognosis of TNBC for the
   first time in our study. These results demonstrated that these four
   genes might be potential biomarkers for clinical diagnosis of TNBC. Our
   findings highlight the potential significance of SHC4, KCNK5, ABCC11,
   and ABCA12 as potential biomarkers for the clinical diagnosis and
   prognosis assessment of TNBC. These discoveries contribute to a deeper
   understanding of TNBC biology and hold promise for the development of
   targeted therapeutic approaches and precision medicine strategies aimed
   at improving outcomes in TNBC patients.

   Growing evidence underscores the role of tumor-infiltrating lymphocytes
   in this subtype of breast cancer ([128]García-Teijido et al., 2016).
   Among these immune cells, tumor-associated macrophages (TAMs), derived
   from blood monocytes, hold significant prominence as they constitute a
   substantial portion of tumor-infiltrating immune cells and are
   influenced by factors secreted by both tumor cells and the tumor stroma
   ([129]Niu et al., 2016). The presence of TAMs has been linked to
   unfavorable prognosis and aggressive tumor characteristics ([130]Zhang
   et al., 2016; [131]Zhao et al., 2017). As the tumor microenvironment
   emerges as a critical target for cancer immunotherapies, comprehending
   the immune cell composition in TNBC and its implications is of
   paramount importance. In this study, We observed that TAMs were the
   predominant immune cell population in both subtypes, a finding
   consistent with previous research ([132]Huang et al., 2022). M1
   macrophages are recognized for their tumor-killing functions, mediated
   through cancer cell recognition, phagocytosis, and proinflammatory
   cytokine production ([133]Choi et al., 2018; [134]Yang et al., 2020).
   Conversely, M2 macrophages have been associated with promoting tumor
   cell invasion, metastasis, angiogenesis, and facilitating immune system
   evasion ([135]Yu and Di, 2017; [136]Linde et al., 2018). However,
   intriguingly, we noted a distinct pattern of macrophage infiltration
   and polarization between TNBC and non-TNBC. Specifically, TNBC
   exhibited significantly higher levels of macrophage infiltration and M1
   macrophage polarization, whereas M2 macrophage polarization was lower
   in TNBC compared to non-TNBC. Our findings align with previous studies
   indicating that M1 macrophage marker genes are significantly increased
   in TNBC compared to non-TNBC ([137]Pe et al., 2022). This increase may
   be attributed to the typically high inflammatory state of TNBC, which
   promotes M1 macrophage polarization via pro-inflammatory cytokines such
   as TNF-α, IL-6, and IFN-γ. Additionally, TNBC often presents with
   elevated levels of tumor antigens, which can stimulate the host immune
   system and recruit M1 macrophages ([138]Li Y. et al., 2022). This
   immune response can result in a more pro-inflammatory tumor
   microenvironment, enhancing M1 macrophage infiltration. M2
   tumor-associated macrophages are enriched in the TNBC microenvironment,
   secreting anti-inflammatory factors that inhibit T cell activity,
   promote immune escape, and support tumor growth ([139]Chang et al.,
   2024). TAMs are predominant immune infiltrating cells in both TNBC and
   non-TNBC. Thus, targeting TAMs, particularly by maintaining the balance
   between M1 and M2 macrophages, is crucial for inhibiting tumor
   progression and immune evasion. In TNBC, reducing the number of M2
   macrophages, hindering the immune escape of tumor cells, and inhibiting
   tumor progression are potential therapeutic strategies. In this study,
   we observed that M2 macrophages were significantly more abundant in
   non-TNBC than in TNBC. This suggests that the higher presence of M2
   macrophages in non-TNBC could lead to greater immune evasion compared
   to TNBC. In non-TNBC, hormone receptor signaling pathways, such as
   those involving ER or PR, may facilitate M2 macrophage polarization. In
   contrast, TNBC, which lacks these signals, exhibits a relatively lower
   proportion of M2 macrophages. Therefore, the immune suppressive
   environment in non-TNBC, characterized by higher M2 macrophage levels,
   presents an opportunity for treatments aimed at depleting M2
   macrophages or blocking tumor immune escape as a potential targeting
   strategy. These findings highlight the importance of understanding the
   macrophage polarization between M1 and M2 macrophages in TNBC and
   non-TNBC, as it holds significant potential for advancing targeted
   immunotherapies and personalized treatment approaches.

   Following transcription factor enrichment analysis, we identified FOXM1
   as a key transcription factor significantly associated with immune cell
   infiltration in breast cancer. FOXM1 is the sole member of the FOXM
   subfamily, which is a critical transcription factor for both the G1-S
   and the G2-M cell cycle transition ([140]Kalathil et al., 2021).
   Notably, FOXM1 has been implicated in various aspects of cancer
   initiation and progression, playing crucial roles in tumor
   angiogenesis, proliferation, migration, invasion,
   epithelial–mesenchymal transition, metastasis, prevention of premature
   cellular senescence, and chemotherapeutic drug resistance ([141]Halasi
   and Gartel, 2013). The oncogenic role of FOXM1 in inducing cell
   proliferation, motility, invasion, and tumor growth in TNBC tumor
   models has been established ([142]Hamurcu et al., 2019). The mechanisms
   by which FOXM1 promotes tumor growth in TNBC mainly include its
   regulation of the expression and activity of focal adhesion kinase
   (FAK) in TNBC cells ([143]Hamurcu et al., 2017), and its direct binding
   to the promoter of the KIF23 gene to promote its transcription and
   accelerate TNBC progression via the Wnt/β-catenin pathway ([144]Li Z.
   et al., 2022). Because the upregulation of FOXM1 to high levels is
   particularly common in TNBC, there is potential for reducing TNBC
   aggressiveness and metastasis by inhibiting FOXM1 activity. Compounds
   that inhibit FOXM1 have been shown to suppress TNBC progression and
   tumor metastasis ([145]Dey et al., 2020). New inhibitors of FOXM1 have
   been highlighted as an attractive target for controlling drug-resistant
   and difficult-to-treat breast cancers ([146]Dey et al., 2020). In
   previous studies, the macrophage-specific deletion of FOXM1 was found
   to reduce the expression of iNOS, IL-1β, and IL-6 ([147]Balli et al.,
   2012). Additionally, FOXM1 inhibition in diabetic mouse models was
   shown to reduce neutrophil and macrophage recruitment to diabetic
   wounds in vivo ([148]Sawaya et al., 2020). These findings indicated
   that FOXM1 activation could increase the expression of pro-inflammatory
   cytokines, driving macrophages toward an M1-like polarization. Our
   study provides additional support for the involvement of FOXM1 in
   macrophage infiltration and polarization in the context of breast
   cancer, particularly between TNBC and non-TNBC. By reducing FOXM1
   expression, there is a potential decrease in the proportion of M1
   macrophages and an increase in the proportion of M2 macrophages. These
   results suggests that in the future research on FOXM1 inhibitors, we
   also need to consider the ratio of M1 and M2 macrophages in vivo to
   maintain the balance of TAMs in the tumor microenvironment, which may
   help develop an attractive drug for controlling drug-resistant and
   difficult-to-suppress breast cancers.

   In summary, this study offers a comprehensive view of immune
   infiltration differences between non-TNBC and TNBC. However, our study
   has several limitations. Firstly, the data used is sourced from the
   [149]GSE76275, which contains a limited number of datasets, potentially
   leading to bias. Secondly, there may be unaccounted confounding
   factors, such as age, sex, or treatment history, in the prognostic
   analysis. Thirdly, the potential biomarkers identified should be
   validated in vitro. Further validation of the biomarkers’ mechanisms
   may facilitate their application in predictive diagnostics, patient
   stratification, targeted prevention, and the personalization of medical
   services for breast cancer.

Conclusions

   Our results indicated that gene co-expression modules 5, 11 and 12 are
   most significantly associated with the subtypes of breast cancer. We
   found that modules 5, 11, and 12 exhibited strong associations with the
   subtypes of breast cancer, with modules 11 and 12 being notably
   enriched in the WNT signaling pathway and lipid/fatty acid metabolism
   pathway, respectively. Within these modules, SHC4, KCNK5, ABCC11, and
   ABCA12 emerged as key genes with significant relevance to the
   progression of TNBC. These key genes proved to be crucial in accurately
   distinguishing between TNBC and non-TNBC, as evidenced by the promising
   average AUC value of 0.963. Of these key genes, KCNK5, ABCC11 and
   ABCA12 genes have a significant impact the prognosis of TNBC. Moreover,
   we observed distinctive immune cell compositions between TNBC and
   non-TNBC, characterized by higher proportions of M0 and M1 macrophages
   in TNBC and lower proportions of M2 macrophages compared to non-TNBC. A
   notable discovery from our study was the identification of FOXM1 as a
   crucial transcription factor significantly associated with macrophage
   infiltration and polarization in TNBC progression. In the future, we
   could develop new FOXM1 inhibitors to inhibit the progression of TNBC
   from the aspects of FOXM1 inhibition and M1 and M2 macrophage balance.
   We also plan to collect clinical triple-negative breast cancer data for
   subsequent validation of these specific markers. These findings provide
   valuable insights into the molecular and immunological aspects of TNBC,
   potentially serving as a basis for the development of new therapeutic
   strategies.

Acknowledgments