Abstract
Background/Aim: Studies on the differences in gene expression characteristics according to lymph node metastasis in breast carcinoma are lacking. This study aimed to compare the spatially resolved transcriptomic profiles between node-positive and negative breast carcinomas. Patients and Methods: Eight patients with breast carcinoma were included, and digital spatial profiling and bioinformatic analysis were conducted to investigate the spatial transcriptomes. Results: In the epithelial compartment, the top two most up-regulated genes were nuclear receptor subfamily 4 group A member 1 (NR4A1) and Jun proto-oncogene, activating protein-1 transcription factor subunit (JUN). The gene ontology (GO) enrichment analysis of the node-positive group revealed a significant up-regulation of genes associated with myeloid differentiation, mononuclear cell differentiation, and hematopoietic regulation. The gene set enrichment analysis (GSEA) revealed significant enrichment of gene sets associated with the regulation of inflammatory cytokines in both the epithelial and stromal compartments of the node-positive group. Conclusion: Our study highlights significant differences in gene expression profiles and spatially resolved transcriptional activities between node-positive and negative breast carcinomas. These findings underscore the importance of developing personalized treatment strategies for lymph node metastasis of breast carcinoma.
Breast carcinoma is one of the most commonly diagnosed malignancies worldwide, posing a significant health burden for women (1-4). Despite advancements in treatment strategies, breast carcinoma remains a leading cause of both incidence and mortality, followed by other cancers such as colorectal and lung carcinomas in incidence rates (5). Notably, even with successful curative surgery with or without postoperative radiation therapy for early-stage breast carcinomas, disease recurrence remains a considerable challenge (6), with approximately 25% of patients with early-stage breast carcinoma experiencing metastatic recurrence within months to decades after initial treatment (6, 7).
Lymph node metastasis (LNM) occurs when tumor cells spread from the primary site to nearby lymph nodes, giving rise to new tumors (8). Consequently, lymph nodes serve as major conduits for disseminating tumor cells to other tissue compartments, thus promoting metastatic cellular proliferation and tumor growth (9). LNM in breast carcinoma is a crucial indicator of advanced stage and poor prognosis (10). Axillary LNM typically represents the earliest detectable clinical manifestation of an emerging distant metastasis (11, 12). Patients with metastatic tumor cells in the axillary lymph nodes exhibit significantly worse prognoses compared to those without LNM (8, 10). Understanding the characteristics and mechanisms underlying axillary LNM is imperative for effectively treating patients with breast carcinoma.
The tumor microenvironment (TME) plays a pivotal role in tumor invasion and metastasis in breast carcinoma (13, 14). The TME encompasses the cellular and molecular components surrounding tumor tissue (15). The tumor cells survive and proliferate by altering TME and reprogramming the surrounding cells, and the interaction between the tumor and immune cells is essential to breast carcinoma metastasis (16).
Digital spatial profiling (DSP) enables multiplexed spatial profiling of RNA using formalin-fixed paraffin-embedded (FFPE) tissue samples (17). Through DSP, RNA transcripts are quantified using ultraviolet (UV)-photocleavable oligonucleotides tagged with antibodies or RNA probes via a UV-linker (17). Light projected onto a tissue sample is photocleaved to release photocleavable oligonucleotides in any spatial pattern over a region of interest (ROI), encompassing 1-5,000 cells (17), providing single-cell sensitivity within an ROI and detecting up to 600 individual mRNA transcripts (17). Therefore, DSP facilitates analyzing the transcriptome characteristics in spatially distinct regions.
Several studies have analyzed the TME of breast carcinoma using spatial transcriptomics (18-24). Liu et al. (18) revealed reduced cytotoxicity and proliferative activity of T cells in metastatic lymph nodes compared with those in breast lesions, by comparing the TME of primary and nodal metastatic tumors using single-cell profiling. Furthermore, mature dendritic cells in nodal metastatic tumors exhibited lower T cell priming and activation capabilities than those in primary tumors (18). Xu et al. (23) used single-cell RNA sequencing and spatial transcriptomics to investigate the transcriptional profile of nodal metastatic breast carcinoma tissues and revealed that tumor-associated macrophages (TAMs) scattered among metastatic tumor cells were enriched in pro-tumoral signaling pathways.
Notably, studies comparing the gene expression characteristics between node-positive and negative breast carcinomas are unavailable. Consequently, this study was aimed at analyzing the spatially resolved transcriptomic profiles of node-positive breast carcinomas using DSP. Our findings may significantly contribute to the understanding of genetic perspectives on breast carcinoma metastasis.
Patients and Methods
Case selection and clinicopathological data collection. The study protocol (2024-03-021) was approved by the Institutional Review Board of Kangbuk Samsung Hospital. Eight patients diagnosed with invasive breast carcinoma between May 2020 and July 2022 were categorized into two groups – the node-positive group and node-negative group – including four patients each with and without LNM, respectively. Electronic medical records were reviewed to collect the following clinicopathological data: patient age, tumor size, LNM, hormone receptor status, Ki-67 labeling index, histological type and grade, and Oncotype Dx Recurrence Score (ORS). Immunoreactivity for hormone receptors and Ki-67 was assessed as described previously (25-30).
Tissue microarray (TMA) technique. TMA blocks were constructed as previously described (31, 32). Briefly, the most representative tumor area, comprising >80% tumor volume, was delineated on the FFPE blocks. Two 3-mm-diameter tissue cores were then extracted from the demarcated area and manually arranged into perforations in the recipient block. Each case was represented by a pair of TMA blocks. DSP technique. Circular ROIs were selected within each TMA core based on the expression of pan-cytokeratin (CK) and smooth muscle actin (SMA), representing the epithelial and stromal compartments, respectively. As shown in Figure 1, these ROIs were subsequently segmented into CK- and SMA-positive clusters. UV exposure of each ROI liberated the barcodes from the DSP probes, which were quantified using a NanoString nCounter system (NanoString Technologies, Seattle, WA, USA). Quality control (QC) and normalization were performed using the GeoMx DSP Analysis Suite (NanoString Technologies). QC assessments included evaluating the binding density of probe tags, performing positive control normalization, and establishing minimum thresholds for nuclei and surface area counts. Specifically, the QC cutoffs applied were a minimum AOINucleiCount of 100, StitchedReads exceeding 1,000,000, and Sequencing Saturation of ≥80%. Normalization was achieved using three housekeeping proteins (histone H3, S6, and glyceraldehyde 3-phosphate dehydrogenase), and relative protein expression data were separately obtained from both compartments. DSP analysis employing the NanoString Whole Transcriptome Atlas (NanoString Technologies) was conducted for each compartment.
Regions of interest obtained using multilabel immunofluorescence staining of invasive breast carcinoma. (A-C) Case 1. (D-F) Case 3. (G-I) Case 5. (J-L) Case 7. The epithelial and stromal compartments are labeled with pan-cytokeratin (pan-CK; green) and smooth muscle actin (SMA; red), respectively.
Bioinformatic analysis. All statistical analyses were performed using R software (version 4.3.1; R Foundation for Statistical Computing, Vienna, Austria). Differentially expressed gene (DEG) analysis was performed using the DESeq2 package (https://bioconductor.org/packages/release/bioc/html/DESeq2.html), with raw data processed to identify DEGs using a threshold of absolute log2 fold change (FC) of >1.0 and an adjusted p-Value of <0.05. The ComplexHeatmap package (https://bioconductor.org/packages/release/bioc/html/ComplexHeatmap.html) was used to generate a heat map of the DEGs, with data scaled by the Z-score for enhanced visualization. Clustering was performed for both rows (representing genes) and columns (representing samples) to identify patterns. The EnhancedVolcano package (https://bioconductor.org/packages/release/bioc/html/EnhancedVolcano.html) was used to generate a volcano plot, juxtaposing statistical significance against the magnitude of FC, enabling rapid identification of genes with significant and large FCs. Gene ontology (GO) enrichment analysis was used to identify enriched GO terms using the enrichGO function from the clusterProfiler package (https://www.bioconductor.org/packages//release/bioc/html/clusterProfiler.html) and org.Hs.eg.db from the OrgDb package (https://bioconductor.org/packages/release/data/annotation/html/org.Hs.eg.db.html), with significance set at p<0.05. Gene set enrichment analysis (GSEA) was used to assess the enrichment of specific gene sets in our dataset using the fgsea package (https://www.bioconductor.org/packages/release/bioc/html/fgsea.html) and MSigDB hallmark gene sets. An adjusted p<0.05 served as the cutoff value for assessing the statistical significance of the estimates.
Results
Table I presents the baseline characteristics of the eight patients with breast carcinoma. All patients were aged >50 years (range=56-67 years), and three patients exhibited single LNM in the node-positive group. All patients were diagnosed with invasive carcinomas of no special type (IC-NST). The mean tumor size was 1.6 cm (range=1.2-1.8 cm). Positive immunoreactivity for hormone receptors and Ki-67 labeling indices was <14% (range=4.2-12.9%) in all cases. Additionally, all patients exhibited a low-to-intermediate risk of recurrence based on the ORS (range=8-20).
Baseline patient clinicopathological characteristics.
Table II summarizes the gene expression profiles of the epithelial compartment. A total of 42 significant DEGs, 31 up-regulated and 11 down-regulated, were identified. Notably, nuclear receptor subfamily 4 group A member 1 (NR4A1), Jun proto-oncogene activating protein-1 transcription factor subunit (JUN), and cysteine-rich 61-connective tissue growth factor-nephroblastoma-overexpressing 1 (CCN1) exhibited the highest FCs (2.3090, 2.1307, and 1.8687, respectively), and their average expression levels were 6.8405, 6.2623, and 5.8106, respectively, in the node-positive group and 4.8586, 4.0960, and 4.3993, respectively, in the node-negative group. In contrast, quinoid dihydropteridine reductase (QDPR), low-density lipoprotein receptor-related protein 5 (LRP5), and stearoyl-coenzyme A desaturase (SCD) were the top three down-regulated genes in the node-positive group, with FCs of −2.0496, −2.0253, and −1.9928, respectively, while their average expression levels were 4.3091, 4.3961, and 4.7524, respectively, in the node-positive group and 6.1713, 5.9534, and 6.5527, respectively, in the node-negative group.
Differentially expressed genes in the epithelial compartment of node-positive breast carcinoma.
Table III presents the gene expression status of the stromal compartment. Domain-containing protein 1 B (DENND1B), keratin 8, and H1.3 linker histone, along with cluster members, were significantly up-regulated in the node-positive group, exhibiting FCs of 2.1831, 1.7165, and 1.5744, respectively, while their average expression levels were 5.6962, 7.1438, and 5.5816, respectively, in the node-positive group and 4.1285, 5.7158, and 4.1527, respectively, in the node-negative group. In contrast, collagen type IV alpha 1 chain (COL4A1), collagen type IV alpha 2 chain (COL4A2), transmembrane 4 L six family member 1 (TM4SF1), and actin alpha 2, smooth muscle (ACTA2) were the most down-regulated genes in the node-positive group, with FCs of – 1.6940, −1.4322, −1.3767, and −1.3186, respectively, while their average expression levels were 5.3920, 5.5233, 4.5248, and 6.6859, respectively, in the node-positive group and 7.1106, 7.0078, 5.8926, and 8.0745, respectively, in the node-negative group. Figure 2 and Figure 3 illustrate the heatmaps and volcano plots of the DEGs.
Differentially expressed genes in the stromal compartment of node-positive breast carcinoma.
Heatmaps of differentially expressed genes in the (A) epithelial and (B) stromal compartments based on lymph node metastasis. The normalized expression values are presented as z-scores to enhance comparative analysis, with the color gradients indicating the degree of expression (red for high expression and violet for low expression). The epithelial compartment heatmap was generated using genes with an absolute fold change (aFC) of >1 and adjusted p-Value of <0.05. The stromal compartment heatmap was generated using genes with an aFC of >1 and a p-Value of <0.01.
Volcano plots of differentially expressed genes in the (A) epithelial and (B) stromal compartments based on lymph node metastasis. Each dot represents a gene, and its location indicates the fold change (FC) of gene expression and its statistical significance. Genes with an absolute FC (aFC) >1 and p-Value of <0.05 are shown for both compartments. The gray dots represent non-significant (NS) genes. The green dots represent genes with significant FC only. The blue dots represent genes with significant p-Value only. The red dots represent genes with significant FC and p-Value.
GO enrichment analysis was conducted to elucidate the functional implications of the identified genes in each compartment of the node-positive group. Notably, significant enrichment of various GO terms related to cellular components, biological processes, and molecular functions was observed. As shown in Table IV, the most significantly enriched terms with counts of ≥10 in the epithelial compartment of node-positive group were associated with the regulation of hematopoiesis (count=11; p<0.0001) and myeloid cell differentiation (count=10; p<0.0001). Furthermore, terms associated with the regulation of leukocyte differentiation (count=9, p<0.0001) and mononuclear cell differentiation (count=9, p<0.0001) were also enriched. In contrast, minimal enrichment was observed in the stromal compartment of node-positive group (Table V), with a count ≤8.
Enriched gene ontology categories of up-regulated genes in the epithelial compartment of node-positive breast carcinoma.
Enriched gene ontology categories of up-regulated genes in the stromal compartment of node-positive breast carcinoma.
Table VI presents the top-enriched pathways for both up-regulated and down-regulated hallmark genes in node-positive group. GSEA revealed several hallmark gene sets that were significantly altered in the node-positive group compared with the node-negative group. Notably, hallmark gene sets associated with the regulation of inflammatory cytokines, such as nuclear factor-B/tumor necrosis factor α (TNF-α) signaling (normalized enrichment score=2.4425 and 2.4909, respectively) and interferon-γ response (normalized enrichment score=1.8815 and 2.2738, respectively) pathways, were significantly up-regulated in both the epithelial and stromal compartments. This observation aligns with the up-regulation of NR4A1, JUN, and CCN1 in the epithelial compartment, and superoxide dismutase 2 (SOD2), TNF-α-induced protein 3 (TNFAIP3), and E74-like E26 transformation-specific transcription factor 3 in the stromal compartment. Additionally, both the epithelial and stromal compartments exhibited significant up-regulation of the hallmark gene associated with the early (normalized enrichment score=1.9181 and 2.3393, respectively) and late (normalized enrichment score=1.8650 and 2.2998, respectively) estrogen response pathways in the node-positive group. In contrast, hallmark gene sets related to the cholesterol homeostasis pathway (adjusted p=0.0186; normalized enrichment score=−1.5796), featuring key genes, such as SCD, clusterin, and fatty acid desaturase 2, were significantly down-regulated in the epithelial compartment of the node-positive group. GSEA of the stromal compartment further revealed negative enrichment of pathways related to the myogenesis (adjusted p=0.0130; normalized enrichment score=−2.1879), epithelial-to-mesenchymal transition (adjusted p=0.0131; normalized enrichment score=−2.1331), and angiogenesis (adjusted p=0.0270; normalized enrichment score=−1.8325). Figure 4 and Figure 5 present the GSEA results for the hallmark pathways significantly up-regulated or down-regulated in the epithelial and stromal compartments of the node-positive group, respectively.
Top enriched hallmark gene sets in the epithelial and stromal compartments of node-positive breast carcinoma.
Gene set enrichment analysis (GSEA) of hallmark pathways in the epithelial compartment. (A) Pink and blue bars represent up-regulated and down-regulated gene sets in the node-positive group, respectively. The length of each bar indicates the normalized enrichment score. (B, C) Most (B) up-regulated and (C) down-regulated gene sets in the node-positive group. The bars at the bottom and top of the panels represent the corresponding genes of certain gene sets.
Gene set enrichment analysis of hallmark pathways in the stromal compartment. (A) Pink and blue bars represent up-regulated and down-regulated gene sets in the node-positive group, respectively. The length of each bar indicates the normalized enrichment score. (B, C) Most (B) up-regulated and (C) down-regulated gene sets in the node-positive group. The bars at the bottom and top of the panels represent the corresponding genes of certain gene sets.
Discussion
For the first time, we analyzed the differences in spatially resolved transcriptomic profiles between node-positive and negative breast carcinoma tissues. Our observation of significantly altered gene expression profiles in node-positive tumors suggests that recognizing the genetic characteristics of LNM may help clinicians develop treatment strategies for advanced-stage breast carcinoma.
NR4A1 and JUN were the most up-regulated genes in the epithelial compartments of the node-positive group. NR4A1 regulates tumor cell survival, migration, and invasion (33), and its overexpression is associated with poor prognosis in several malignancies (34-41). Zhou et al. reported that NR4A1 is highly expressed in breast carcinoma, exhibits high immune cell infiltration, and is associated with poor patient prognosis (34). Furthermore, NR4A1 induces an inflammatory response, promoting the invasion and metastasis of breast carcinoma cells (34). JUN is an oncogene involved in breast carcinoma (42, 43), and it is highly expressed at the invasive tumor front and is associated with tumor cell proliferation, invasion, migration, and angiogenesis (42, 44). Up-regulation of NR4A1 and JUN in the node-positive group in this study indicates their involvement in the metastatic progression of breast carcinoma.
From a therapeutic standpoint, conflicting data exist regarding the roles of NR4A1 and JUN in tamoxifen resistance. Kim et al. (35) reported that enhanced NR4A1 expression improved responsiveness to tamoxifen in patients with ER-positive breast carcinoma who developed tamoxifen resistance. Furthermore, NR4A1 inhibited extracellular signal-regulated kinase signaling in tumor cells, thereby increasing drug responsiveness (35), suggesting that increasing NR4A1 could be a potential therapeutic target to overcome tamoxifen resistance in ER-positive breast carcinomas. In contrast, He et al. (43) reported that overexpression of JUN, an AP-1 transcription factor, reduces sensitivity to tamoxifen in patients with ER-positive breast carcinoma, suggesting that inhibiting JUN might improve tamoxifen responsiveness in AP-1-overexpressing ER-positive breast carcinomas (45). The role of NR4A family members in T cell responses has also been described. Deficiencies of NR4A1 and NR4A2 in T cells enhance their antitumor activity (46), indicating NR4A family targeting as a promising therapeutic approach to boost antitumor immunity. Srirat et al. (46) reported NR4A1 and NR4A2 inhibition as a powerful immunotherapeutic strategy to reduce the exhaustion of cytotoxic T cells. Notably, analogs of bis-indole-derived compounds that bind to and antagonize NR4A1 have emerged as potential therapeutic agents for breast carcinoma (47).
In the stromal compartment of the node-positive group, SOD2 and TNFAIP3 were significantly up-regulated. Numerous studies have reported elevated SOD2 expression in the early stages of tumor growth and during metastatic progression (48-51), supporting the oncogenic role of SOD2 in breast carcinoma (52, 53). Moreover, elevated SOD2 expression facilitates invasion and metastasis of breast carcinoma by enhancing the activity of cellular oxidants (48, 50). Ennen et al. (52) reported that high SOD2 levels enhance the invasive and metastatic abilities of tumor cells. Notably, Li et al. (54) reported an association between SOD2 overexpression and axillary LNM. However, all these previous studies have focused exclusively on SOD2 expression in the tumor cells but not in the TME, and only a single study by Trimmer et al. is available on SOD2 expression in the stromal compartment (55). Consistent with our data, Trimmer et al. reported that SOD2 is transcriptionally up-regulated in the stromal compartment of breast carcinoma tissue and that SOD2 overexpression reduced the tumor-promoting effects of tumor-associated fibroblasts, suggesting a potential stromal tumor suppressor role for SOD2 in breast carcinomas (55). Nevertheless, further investigations are warranted to elucidate the expression and functional significance of SOD2 within stromal cells in the TME of breast carcinoma.
TNFAIP3 is an anti-inflammatory protein that protects tissues against TNFα-induced cytotoxicity by limiting the NF-B signaling pathway, and its abnormal expression in various tumors indicates its essential role in tumor development and progression (56). The mechanism underlying aberrant TNFAIP3 expression is associated with anti- or pro-apoptotic regulation, dependent mainly on cell-specific and apoptosis-stimulating factors (56). Although TNFAIP3 expression decreases in pancreatic and colorectal carcinomas (57, 58), its increased expression has been reported in carcinomas of the hepatobiliary tract and esophagus (59-61). In ER-positive breast carcinoma, TNFAIP3 overexpression protects tumor cells from TNFα-induced cell death (62). However, drawing conclusions about the effect of TNFAIP3 overexpression in the stromal compartment is challenging, as studies examining the stromal expression of TNFAIP3 in breast carcinoma remain lacking.
We observed up-regulation of genes associated with myeloid and mononuclear leukocyte differentiation and hematopoiesis regulation in the epithelial compartment of the node-positive group. Additionally, genes involved in adaptive immune responses, lymphocyte-mediated immunity, and leukocyte-mediated immunity were up-regulated in the stromal compartment of the node-positive group. Myeloid and mononuclear cells play crucial roles in the TME. Various human malignancies are associated with the expansion of immunosuppressive cells, such as myeloid-derived suppressor cells (MDSCs) and TAMs (63), which promote tumor cell growth through various mechanisms (64). The accumulation of MDSCs in the TME diminishes the natural antitumor abilities of natural killer cells and T cells (64). Domagala et al. (65) reported that tumor-infiltrating myeloid cells contribute to immune evasion, tumor growth, migration, and resistance to treatment. Moreover, Xu et al. reported that TAMs activate the pro-oncogenic pathway via NF-B signaling in metastatic lymph nodes (23). Similarly, Mao et al. (19) demonstrated that TAMs expressing both M1 and M2 signatures exhibit high tumor specificity, and the immune milieu in metastatic lymph nodes transforms into a tumor-like state, characterized by increased pro-inflammatory macrophages and exhausted T cells (19). Tumor-associated hematopoietic dysregulation results in myeloproliferation and host immunity suppression (63). In particular, the TME modulates myeloid cells into potent immunosuppressive cells (66). Our observations of enriched hallmark gene sets related to inflammatory cytokines, including NF-
B/TNF-α signaling and interferon-γ response pathways in both the epithelial and stromal compartments, support the notion that tumor metastasis is sustained by numerous altered molecular pathways involving various biological crosstalk among different cellular components within the TME (67).
This study had three principal limitations. Firstly, the sample in this study was small, and only eight patients with grade 1-2 ER-positive IC-NST participated in this study, four of whom experienced LNM. Consequently, a small sample size limits the generalizability of our findings, and our results may not adequately represent the diversity of perspectives. Furthermore, we could not examine whether the differences in altered gene expression levels and pathways have a significant prognostic value for predicting clinical outcomes owing to the limited sample size. Secondly, this study was conducted at a single institution, limiting the reproducibility of the results elsewhere. The primary drawback of single-institution studies is their limited external validity. Future investigations with larger multi-institutional cohorts of patients with breast carcinoma are warranted to validate our findings and enhance the reliability of LNM-related differences in spatially resolved transcriptomic profiles. Thirdly, we could not examine matched primary and nodal metastatic breast carcinoma tissue samples to investigate the molecular concordances or discrepancies between the two samples.
Conclusion
We demonstrated that node-positive breast carcinoma exhibits significantly different gene expression profiles and spatially resolved transcriptional activities compared with node-negative breast carcinoma. We identified significantly up-regulated or down-regulated genes, hallmark gene sets, and pathways associated with the invasion, migration, and metastasis of tumor cells, regulation of the immune response, and drug responsiveness. Our findings emphasize the need for personalized treatment strategies for patients with breast carcinoma.
Acknowledgements
This work was supported by a Samsung Medical Center Grant (SMO1240641) and a National Research Foundation of Korea (NRF) grant funded by the Korean Government (MSIT) (2023R1A2C2006223).
Footnotes
Authors’ Contributions
All Authors made substantial contributions to the conceptualization and design of this study; collection, interpretation, and validation of the data; drafting of the manuscript; critical revision of the manuscript; and approval of the final version to be published.
Conflicts of Interest
None of the Authors have any conflicts of interest or financial ties to declare regarding this study.
- Received May 13, 2024.
- Revision received June 11, 2024.
- Accepted June 17, 2024.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).