Estradiol-associated extracellular signaling in the human breast in vivo identifies mast cells as tumor-supportive stromal mediators Authors: Annelie Abrahamsson, Xingyue Wang, Wen Zhong, Charlotta Dabrosin Published: 2026 ABSTRACT AbstractSex steroid exposure is a major risk factor for breast cancer initiation and progression through incompletely understood mechanisms. Novel therapeutics for prevention and treatment of estrogen-dependent breast cancer are needed. Cell-cell interactions within the microenvironment contribute to breast cancer progression through extracellular signaling. Here, extracellular proteins were sampled directly from live normal human breast tissue and breast cancers using microdialysis, and 1,101 proteins were quantified. Mechanistic studies were performed in cultured intact normal human breast tissue and mast cells. Findings were validated in vivo using ER+ breast cancer models in immunocompetent and immunocompromised mice and in a zebrafish xenograft model. We identified a breast tissue-specific extracellular protein signature comprising 150 proteins associated with local estradiol (E2) levels, of which several were linked to mast cell biology, including stem cell factor (SCF). Human ER+ breast tumors exhibited increased extracellular SCF levels and higher mast cell density, which was associated with poorer overall survival. In normal human breast tissue, E2 increased mast cell abundance and activity. In vitro, E2 rapidly activated mast cells, an effect blocked by desloratadine. In mouse breast cancer models, E2 increased and fulvestrant reduced mast cell density. In the zebrafish, mast cells promoted ER+ breast cancer growth, which was further enhanced by E2 and abrogated by desloratadine. Overall, our findings identify mast cells as estrogen-responsive stromal mediators and support targeting mast cell-dependent pathways as a potential strategy for the prevention and treatment of estrogen-dependent breast cancer. Introduction Cumulative exposure to sex steroids constitutes an independent risk factor for sporadic breast cancer, which accounts for over 90% of the cases, and reduction of estrogen signaling with tamoxifen or aromatase inhibitors can substantially reduce this risk. 1-3 However, adverse effects, including endometrial cancer, thromboembolic events and sexual dysfunction, leading to poor adherence, in combination with incomplete protection, limit their use as preventive strategies. 4 , 5 Identification of additional estrogen-regulated pathways requires a deeper understanding of the human breast microenvironment and the role of estrogen in its regulation. Estrogen signaling in the breast extends beyond the epithelial compartment, which represents less than 10% of the tissue. 6 The surrounding stroma contains multiple cell populations, including immune cells, and stromal activation has been shown to be a prerequisite for tumor progression. 7-9 We have previously shown that estradiol (E2) modulates innate immune cells, including macrophages and neutrophils, within the human breast microenvironment. 10-13 Whether mast cells constitute another estrogen-responsive component of the breast microenvironment has not been established. Mast cells participate in mammary gland development and tissue remodeling, whereas their role in breast cancer progression appears to differ between molecular subgroups. 14 , 15 However, whether estrogen directly regulates mast cell function in the human breast and contributes to tumor-promoting stromal remodeling remains unclear. In this study, we investigated how estrogen shapes the extracellular microenvironment of the normal human breast. We combined in vivo microdialysis of breast tissue from healthy pre- and postmenopausal women with high-dimensional proteomic profiling, using their marked difference in endogenous estrogen exposure to identify breast-specific E2-associated extracellular signaling. In an unbiased screen, a prominent mast cell–associated signature was revealed, which was further investigated in human breast tissue culture, cell culture, and in vivo experimental models. Finally, we examined the relevance of this pathway in human ER+ breast cancer. Materials and methods Study participants The Regional Ethical Review Board of Linköping approved all studies, including biobanking of tissues (#2011/473-32, #2014/185-32, #M61-05/32-06, for breast cancer patients, women with nondense or dense breasts, and premenopausal women, respectively). All studies were carried out in accordance with the Declaration of Helsinki. All subjects gave informed written consent. Collected and biobanked samples from different cohorts were used in the exploratory clinical studies. Parts of these cohorts and microdialysis samples have been used in previous analyzes addressing different biological questions. 11 , 16-20 A total of 52 women were included. Nineteen nulliparous women (age: 20–32 years) with a history of regular menstrual cycles (cycle length: 27–34 days) were included in the healthy premenopausal group. 16 The premenopausal women underwent microdialysis during the luteal phase of the menstrual cycle. As previously described, healthy postmenopausal women (age: 55–74 years) were recruited from the mammography screening program at Linköping University Hospital, where their breasts were categorized according to the Breast Imaging Reporting and Data System (BI-RADS) as either entirely fatty nondense (BI-RADS A) or extremely dense (BI-RADS D). 11 , 20 Because premenopausal women have dense breast tissue per se, only postmenopausal women with dense breasts (n = 21) were included in the analyzes to avoid skewing of the results when comparing pre- and postmenopausal breast tissue. 21 None of the healthy participants had a history of breast cancer or were currently using (or had used within the past 3 months) any steroid-containing medications. Additionally, we included 12 postmenopausal women (age, 52–86 years) diagnosed with ER+ breast cancer who underwent microdialysis the day before surgery as previously reported. 17 Microdialysis procedure Microdialysis was performed as previously described. 12 , 17 , 22 , 23 In brief, catheters equipped with 100-kDa cut-off membranes were perfused at 0.5 μL/min. In healthy participants, 20-mm membranes were positioned in breast tissue and abdominal subcutaneous fat. In patients with breast cancer, 10-mm membranes were placed within the tumor and in adjacent normal breast tissue. Following a 60-min equilibration period, microdialysates were collected and stored at −80 °C until analysis. Protein quantification and estradiol analysis of microdialysates Proteins in microdialysates were quantified using multiplex proximity extension assay (PEA; Olink Bioscience, Uppsala, Sweden), as previously described. 10 , 16 , 17 , 19 , 24 The analysis comprised 1,193 proteins, representing 1,158 unique proteins across 13 target panels (Olink Bioscience, Uppsala, Sweden). Pre-processed data were provided in arbitrary normalized protein expression (NPX) units on a log2 scale and linearized using 2NPX transformation. The values thus represent a relative quantification, meaning that no comparison of absolute levels between different proteins could be made. E2 in the microdialysates was analyzed using a high-sensitivity immunoassay kit (DRG International, Springfield Township, NJ, U.S.A.). Of the 1,158 unique proteins analyzed in the present study, 92 were also included among the 461 proteins analyzed using a different proteomic panel in our recent study of inflammatory breast tissue microenvironments associated with mammographic density and estrogen exposure. 16 Intact normal human breast tissue culture Biobanked cultured samples from reduction mammoplasties from five donors along with their corresponding culture media were analyzed. Biopsies containing epithelium, stroma, and adipose tissue were obtained using an 8 mm biopsy punch (Kai Europe GmbH, Solingen, Germany), placed individually on 12-well plates (Costar, Cambridge, Mass. U.S.A. Cat# 3513), and incubated at 37°C in a humidified atmosphere with 5% CO2 in serum-free medium ± 1 nM E2 (17β-estradiol; Sigma, Cat# 275817) as previously described. 25 After seven days, media were collected and biopsies were fixed in 4% buffered formalin. Within each experiment, several separate tissue biopsies were allocated to control or E2 treatment and served as replicate samples. Measurements from replicate biopsies, microscopic fields, or individual cells are presented as medians within each independent experiment before statistical analysis. Cell lines and culture conditions ER+ breast cancer cells, MCF-7 (RRID:CVCL_0031) were maintained in Dulbecco´s modified Eagle’s medium (DMEM) (Gibco, Cat# 11880) with 2 mM glutamine (Gibco, Cat# 25030), 50 IU/ml penicillin-G, 50 mg/ml streptomycin (Gibco, Cat# 15070), and 10% fetal bovine serum (FBS, Gibco Cat# 10270). LUVA cells (Kerafast, Boston, MA, U.S.A., Cat# EG1701-FP) were maintained in STEM PRO-34 SFM (Gibco, Cat# 10640) with STEM PRO-34 (Gibco, Cat# 10641-025), 2 mM glutamine (Gibco, Cat# 25030), 50 IU/mL penicillin G, and 50 mg/mL streptomycin (Gibco, Cat# 15070). HMC-1.1 cells (RRID:CVCL_H206), a kind gift from Dr. JD Söderholm, Linköping University, Sweden, were maintained in Iscove's modified Dulbecco's medium (IMDM; Gibco, Cat# 21056023) with 50 IU/mL penicillin-G, 50 mg/mL streptomycin, 1.2 mM α-Thio glycerol (Sigma, St Louis, MO, U.S.A., Cat# M6145), and 10% fetal bovine serum (Gibco Cat# 10270). Cells were diluted in serum-free culture medium, DMEM/F12 (Gibco, Cat# 11039) with 0.02% bovine serum albumin (Merck, Darmstadt, Germany, Cat # 1.12018.0025), 10 μg/mL apo-transferrin (Sigma-Aldrich, Saint Louis, U.S.A., Cat# T2036), and 1 μg/mL insulin (Sigma-Aldrich, Cat# I5500), and seeded in a 24-well plate at 2 × 105 cells/well ± 1 nM E2 and ± 20 μM desloratadine (MedChemExpress, Newark, NJ, U.S.A., Cat# Hy-B0539). Culture medium was analyzed with Human Luminex Discovery kit (Bio-Techne, Minneapolis, USA, Cat# LXSAHM) with Luminex FLEXMAP 3D Instrument System (Luminex Corporation, Austin, Texas, USA). SCF (KIT ligand) and VEGF were quantified using ELISA kits (SCF, Invitrogen, Thermo Scientific, MD, U.S.A., Cat# EHKITLG, VEGF QuantiGlo Bio-Techne, Cat#QVE00B) using a VersaMax™ Microplate Reader (Molecular Devices, Sunnyvale, CA, USA). 4 × 104 cells were seeded in 96-well plates. After 24 h, cells were treated ± 1 nM E2 for 72 h and proliferation was analyzed with BrdU kit (Roche Diagnostics, Mannheim, Germany, Cat# 11 647 229 001) and VersaMax™ Microplate Reader. Mast cells were diluted in serum-free culture medium and 2.5 x 105 cells and seeded on chamber slides (Thermo Fischer Scientific, NY, U.S.A., Cat# 177402), and treated with ± 1 nM E2, ± 1 μM fulvestrant, and ± 20 μM desloratadine (MedChemExpress, NJ, U.S.A., Cat# Hy-B0539) for 5, 20, or 60 min at 37 °C, 5% CO2, and 99% humidity. The concentrations of E2 and desloratadine were selected based on previously published in vitro studies of estrogen signaling in cancer cell and mast cell responses, respectively. 26 , 27 Culture medium was removed and cells were fixed in 4% buffered formalin. ER was verified with rabbit anti-human ER (Biocare, CA, U.S.A., Cat# BC-CRM301C; Leica Biosystems) using the Leica Bond-III system with the Bond Polymer Refine Detection kit (Leica Biosystems, Cat# DS9800) and Olympus BX43 light/fluorescence microscope. Murine breast cancer models The Institutional Animal Ethics Committee of Linköping University approved this study (protocol #127-11 and #67-12, for BALB/c nu/nu and FVB/N respectively), which conformed to the regulatory standards for animal care. At the endpoint of the study, mice were euthanized by cervical dislocation under full anesthesia induced by intraperitoneal injection of ketamine/xylazine in accordance with the American Veterinary Medical Association. Oophorectomized athymic mice, BALB/c nu/nu, and FVB/N mice (6-8 weeks old, Scanbur, Sweden), were housed at Linköping University in ventilated cages with a light/dark cycle of 12/12 h with rodent chow and water available ad libitum. Mice were anesthetized via intraperitoneal injection of ketamine/xylazine and implanted with a subcutaneous (s.c.) 3-mm pellet containing either 17β-estradiol (0.18 mg/60-day release, Innovative Research of America, Sarasota, FL, U.S.A.), resulting in serum concentrations of 150-250 pM estradiol, or placebo. 28 Seven days after surgery, 5 × 106 MCF-7 cells in 200 μL PBS were injected in the dorsal mammary fat pad in mice with estradiol pellets as MCF-7 cells require estrogen for tumor formation and growth. When MCF-7 tumors were established, mice bearing tumors of similar sizes were randomly divided into two groups, where one group was exposed to E2 only (n = 5) and the other to fulvestrant (5 mg/mouse twice per week, s.c.) added to the E2 treatment (n = 5). The FVB/N mice were injected into the dorsal mammary fat pads with 1 × 106 tumor cells, in 200 μL PBS, derived from a transgenic mouse strain expressing polyoma middle T (PyMT) antigen under the control of the mouse mammary tumor virus (MMTV) long terminal repeat. Tumors were established in both E2 treated (n = 5) and placebo mice (n = 5). Group sizes were based on previous studies using these models. 29-31 Immunofluorescence Formalin-fixed normal breast tissue and mouse tumors were paraffin-embedded, cut into 4-μm sections, and human mast cell lines were stained with mouse anti-human mast cell tryptase (Santa Cruz Biotechnology Cat# sc-59587, RRID:AB_793510, 1:200) and goat anti-mouse MCP-4 (Abcam, Cambridge, UK, Cat# ab92368, 1:50) as primary antibodies followed by goat anti-mouse Alexa Fluor 488 nm (Life Technologies Corporation, Eugene, OR, U.S.A., Cat# A11001, 1:200) or donkey anti-goat Alexa Fluor 488 nm (Invitrogen, Cat# A11055, 1:200) as secondary antibodies with Hoechst blue counterstain (ex 460–490 nm; Thermo Fisher Scientific, Cat# 33342, 1:5000). Antibodies were diluted in Antibody Diluent, Background Reducing Reagent (Agilent/DAKO, Glostrup, Denmark, Cat# S302281-2). Sampes were visualized with Olympus BX43 microscope, excitation filters BP360-370 and BP460-495, Olympus DP74 3CCD mode camera, and CellSens imaging software version 1.15 (Olympus CellSens Software, RRID:SCR_014551). In the cell culture experiments, the percentage of cells exhibiting detectable tryptase fluorescence was determined. Cells classified as tryptase-negative showed no detectable fluorescence. In tumor tissue and breast biopsies, fields were selected as mast cell hotspots. The examiner was blinded to treatment groups during analysis. The fluorescence intensity was measured using Fiji (RRID:SCR_002285). Color channels were split into separate grayscale images. A fixed threshold of 36-255 was applied and cells were outlined with the freehand tool, and mean intensity was measured. Immunohistochemistry Sections were stained for mast cell tryptase using a fully automated Leica Bond-III system (Leica Biosystems, Melbourne Pty Ltd, Melbourne, Australia) at the pathology department at Linköping University Hospital with a mononuclear mouse anti-human mast cell tryptase antibody (Cat# PA0019, Leica Biosystems). The Bond Polymer Refine Detection Kit (Leica Biosystems, Cat# DS9800) was used for detection and analyzed with Olympus BX43 light fluorescence microscope. Zebrafish experiments The Linköping Animal Research Ethical Committee approved the experiments (protocol #20014-2023) that were performed by a CRO (BioReperia AB). The TL wild-type strain was purchased from the European Zebrafish Resource Center (Germany) and raised and maintained in a zebrafish facility at Linköping University, Linköping, Sweden, as previously described. 32 Zebrafish embryos were maintained in E3 embryo medium (containing 0.286 g NaCl, 0.048 g CaCl2, 0.081 g MgSO4 and 0.0126 g KCl/L) supplemented with 0.2 mM 1-phenyl-2-thiourea (PTU) and incubated at 28.5 °C. One day before injection, MCF-7 cells were labeled with the Dil dye (VWR, Cat #APOSBIB6289). MCF-7 cells were implanted with ± LUVA cells at a ratio of 70% MCF-7 cells to 30% LUVA into the perivitelline space of 2-day-old zebrafish embryos. Tumor-bearing embryos were randomly sorted and treated with vehicle control, 1 nM E2, or 20 μM desloratadine added to embryo water. Tumor-bearing embryos were incubated at 35.5 °C for 3 days. The embryos were anesthetized using MS-222 (Sigma-Aldrich, Cat #E10521) 0.5 mg/mL and images were acquired on days 0 and 3 using a stereoscope model M205 FA (Leica Microsystems CMS GmbH) equipped with a K5 camera (Leica) with resolution (h × v) 2048 × 2048 pixels and pixel size (h × v) 6.5 × 6.5 μm, at an image magnification of 100x. Images were analyzed using Hugin&Munin software v2.10.1.0 (BioReperia AB, Linköping, Sweden). Relative tumor sizes were analyzed as the area of labeled tumor cells on days 0 and 3 after tumor implantation in the same embryo blinded to the examiner. At the endpoint on day 3, the anesthetized embryos were euthanized by hypothermal shock through exposure to 4 °C for at least 24 h. Following confirmation of the absence of heartbeat and spontaneous movement, embryos were transferred to −20 °C. Breast cancer database analysis The Kaplan-Meier plotter (KMP) database for mRNA sequencing was used to analyze overall survival of patients with ER+ PR+ HER2− breast cancer (https://kmplot.com/analysis/) with settings: auto select best cut-off percentile, follow up threshold all, lymph node status all, Ki67 all, Nottingham histologic grade all, PAM50 subtype all, and all endocrine and chemotherapy treated patients. The analysis was exploratory and univariable; no multivariable adjustment was performed. Statistical analyzes Non-normally distributed data were analyzed using two-sided nonparametric Mann–Whitney U tests for unpaired two-group comparisons, Kruskal-Wallis with Dunn’s post hoc test for multiple-group comparisons, and Wilcoxon’s signed-rank test for paired data. In the first screen the false discovery rate (FDR) was set to 5% using the two-stage method of Benjamini, Krieger, and Yekutieli. In the subsequent analysis of correlations between breast tissue-specific proteins and local breast E2 concentrations, the non-adaptive Benjamini–Hochberg procedure was applied with an FDR of 5% because the tested protein set had already been selected in the preceding differential-abundance analysis. Gene ontology (GO) enrichment analysis (Biological Process), KEGG and Reactome pathway enrichment analysis were performed based on the significant proteins in the expression analysis (P < 0.05). Significant protein-protein associations were identified based on Spearman’s correlation coefficient (R) and P-values, using thresholds of |R| > 0.8 and FDR < 1 x 10−7. The protein-protein correlation network analysis incorporated 235 protein pairs, comprising 100 individual proteins. A weighted and undirected protein-protein network was constructed, with edge weights reflecting the absolute values of the correlation coefficients. The Louvain algorithm was used for protein community analysis. Key centrality metrics, including the degree, betweenness, and closeness, were calculated for each protein. Network analysis and visualization were performed using the R packages igraph v2.0.3 and qgraph v1.9.8. GO and KEGG enrichment analyzes and visualization of protein communities were conducted using R packages clusterProfiler v4.10.1 and ggplot2 v3.5.1, with the 1,158 proteins quantified in the study used as the background (R version 4.3.3). Statistical methods are specified in the figure legends. All statistical tests were performed using Prism 10.0 (GraphPad, San Diego, CA, U.S.A.) and R version 4.2.2 (2022-10-31). Results Characterization of the study participants There was no difference in the body mass index between the pre- and postmenopausal groups (24 ± 1.5 and 24 ± 3.3, respectively). Local breast E2 levels were significantly lower in the postmenopausal group than in the premenopausal group (40 ± 14 pmol/L and 192 ± 50 pmol/L, respectively; P < 0.0001). Systemic and breast tissue extracellular protein profiles Of the 1,158 individual proteins on the panels, 1,101 proteins were detected above the lowest level of detection (LOD) in > 50% of the samples in at least one group. Of these, 551 proteins were significantly changed in premenopausal breast tissue compared to postmenopausal breast tissue (FDR-corrected P < 0.043, Figure 1A). To distinguish local breast tissue alterations from systemic differences, we also sampled proteins from abdominal subcutaneous fat by microdialysis. The protein profile of abdominal subcutaneous fat exhibited a pattern similar to that of breast tissue. After FDR correction (P < 0.029), 440 proteins were significantly altered between the pre- and postmenopausal groups (Figure 1B). Signaling pathways were analyzed using the overall changes in breast tissue without any consideration of breast tissue-specific effects. The results of the reactome pathway analysis and GO enrichment analysis (Biological Process) are shown in Figure 1C–F. Different in vivo extracellular protein signatures systemically and locally in normal breast tissue of pre- and postmenopausal women. Microdialysis for in vivo sampling of local extracellular proteins was performed in the breast tissue and abdominal subcutaneous fat as a proxy for systemic alterations in healthy pre- and postmenopausal women (n = 19 and n = 21, respectively). Proximity extension assay was used for quantification of 1,158 individual proteins in the microdialysates. 1,101 proteins were above the lowest level of detection (LOD) in > 50% of the samples in at least one group. A) Profile of extracellular proteins in premenopausal (high estradiol) versus postmenopausal (low estradiol) normal breast tissues. B) Proteins in the abdominal subcutaneous fat from the same participants as in (A). FDR = False Discovery Rate. C) A network of the Reactome Pathway analysis of altered proteins in pre- and postmenopausal breast tissue. D) Dotplot of two enriched pathways from the reactome pathway analysis. E) A network of the top five enriched terms from Gene Ontology analysis of proteins altered in pre- vs. postmenopausal breast tissue. F) Dotplot of the top 10 Gene Ontology terms. Proteins with breast tissue-specific and systemic differences between pre- and postmenopausal women To identify breast tissue-specific alterations in protein levels between the healthy pre- and postmenopausal groups, we compared the proteins that were altered in breast tissues with those that were altered in abdominal subcutaneous fat. Of the 551 proteins that were altered in breast tissue, 202 were altered only in breast tissue, whereas the other 349 were altered in both breast and abdominal fat tissues. In addition, 125 proteins were altered in abdominal fat tissue only (Supplementary Table 1). Co-expression network of proteins with breast tissue-specific differences between pre- and postmenopausal women The 202 proteins that were altered in breast tissue only were analyzed by Spearman’s correlation analysis. We identified 4,447 protein pairs with FDR < 0.01, of which 95.4% showed positive pairwise correlations (Figure 2A). Additionally, 235 protein pairs involving 100 individual proteins had an FDR < 1 x 10−7 and |R| > 0.8. Enrichment analysis showed that most of these proteins were involved in immune-related pathways (Figure 2B). A protein-protein association network was constructed using the 235 protein pairs identified (Figure 2C). Co-expression of tissue-specific proteins in pre- vs. postmenopausal breasts. Extracellular proteins were sampled using microdialysis, as described under Figure 1. Proteins that were differentially expressed in the breast tissue were identified. A) Barplot showing the number of positive and negative pairwise associations (FDR < 0.01) among the 202 breast tissue-specific proteins comparing pre- and post-menopausal dense breast tissue. The density plot shows the distribution of Spearman’s correlation coefficients, with negative correlations, positive correlations, and all associations colored blue, red, and gray, respectively. B) Barplot showing the results of KEGG enrichment analysis of the 100 proteins with FDR < 1 x 10−7 and an absolute value of Spearman correlation coefficients > 0.8. C) A protein-protein association network based on 235 associations between 100 proteins, with the different communities circled by gray shadows and nodes colored according to the associated community. Node size was positively correlated with betweenness score. Positive correlations are indicated by gray lines, and negative correlations are indicated by blue lines. Correlation between local E2 levels and proteins with breast tissue-specific differences between pre- and postmenopausal women Next, we analyzed the correlations between the 202 proteins with breast tissue-specific differences and local breast tissue E2 levels. A total of 161 breast tissue-specific proteins were significantly correlated with E2 levels with a P < 0.05. After FDR correction, 150 remained significant including 19 proteins with negative correlations and 131 proteins with positive correlations (Figure 3A). The three proteins showing the strongest E2 correlations were CD63, NDRG1, and SCF (Figure 3A). Their extracellular levels in pre- and postmenopausal women are depicted in Figure 3B. Of the 150 proteins with significant correlations with E2, 56 proteins, including CD63, NDRG1, and SCF, were shown to be either secreted or to affect mast cell functions (Supplementary Table 2). Of these 150 proteins, only 17 were included in the proteomic panel used in our previous study of estrogen exposure, 16 whereas 133 (89%) were not analyzed previously. Associations between local breast estradiol (E2) levels and breast tissue-specific proteins. Extracellular proteins were sampled using microdialysis, as described under Figure 1. A) Spearman’s correlation coefficients and corresponding P values for associations between local breast E2 levels and the 202 proteins showing breast tissue-specific differences between pre- and postmenopausal women. The horizontal line indicates the Benjamini–Hochberg FDR threshold. CD63, NDRG1, and SCF showed the strongest positive correlations with E2. B) Extracellular levels of CD63, NDRG1, and SCF in breast tissue from premenopausal (n = 19) and postmenopausal (n = 21) women. Box plots show min–max and all individual data points. Mann–Whitney U test, ****P < 0.0001. E2 increased mast cell numbers and decreased their tryptase immunoreactivity in parallel with increased inflammatory cytokines in cultured intact normal human breast tissue To determine whether E2 exerts effects on mast cells in normal human breast tissue, biobanked cultured intact normal human breast tissue obtained from reduction mammoplasties in the presence or absence of E2 were analyzed. As shown in Figure 4A and 4B, exposure to E2 significantly increased the number of mast cells but reduced intracellular tryptase immunoreactivity, consistent with increased mast cell activation. The increased number of mast cells may reflect enhanced survival of resident mast cells in response to E2. Thereafter, we analyzed the culture media from the breast biopsies. Proteins known to be secreted by mast cells and significantly correlated with E2 in the in vivo data were analyzed, including CCL2, CCL4, CCL23, IL10, IL13, IL17, MIA, SCF, TNFα, and VEGF. CCL2, CCL4, CCL23, TNFα, and VEGF were significantly increased in the media after E2 exposure (Figure 4C). However, the breast biopsies contain multiple cell types, and the cellular origin of these proteins cannot be determined from this experiment. Therefore, we next performed mast cell culture experiments to determine whether E2 can exert direct effects on mast cells. Estradiol (E2) increased the mast cell number and altered mast cell-associated responses in ex vivo culture of intact normal human breast tissue. Intact normal human breast tissue biopsy samples from patients subjected to reduction mammoplasties were cultured in the presence or absence of 1 nM E2 for seven days (n = 5). Breast tissues were stained for mast cells, and the released proteins were analyzed in the culture media. A) Number of tryptase-positive cells (mean of 5 hot spot areas per sample in X400 high-power fields). B) Mean fluorescence intensity of the mast cells in each group detected in Figure 4A. Box plots show min–max and all individual data points. Mann–Whitney U test *P < 0.05, **P < 0.01. Representative sections from each group are shown; scale bar, 20 μm. C) Proteins released in the cell culture media. Box plots show min–max and all individual data points. Mann–Whitney U test, *P < 0.05, **P < 0.01, ****P < 0.0001. E2 induced rapid activation accompanied by cytokine secretion in cultured mast cells To determine whether E2 also exerts any direct effects on mast cell degranulation or activation, we cultured HMC-1.1 cells, which originated from a patient with mast cell leukemia and harbored a mutation in the KIT receptor. In the first set of experiments, we determined whether these cells expressed ER. As shown in Figure 5A, the HMC-1.1 cells were positive for ERα. However, these cells did not proliferate in the presence of E2 for 72 h (Figure 5B). Next, we investigated the possible effects of E2 on granule release from mast cells. As shown in Figure 5C, E2 induced a time-dependent loss of intracellular tryptase immunoreactivity in HMC-1.1 cells; representative cells are depicted in Figure 5C. Desloratadine counteracted the E2-dependent loss of intracellular tryptase immunoreactivity (Figure 5D). To elucidate whether the release of granules was accompanied by cytokine secretion, we analyzed the same panel of proteins detected in human samples. After 20 min, no changes were observed in the released proteins. However, after 60 min, CCL2 and VEGF levels were significantly increased, which was reversed by treatment with desloratadine (Figure 5E). IL13, IL17, and MIA were unchanged after E2 treatment, whereas CCL4, CCL23, IL10, SCF, and TNFα were all below the detection limit in the cell culture media. Estradiol (E2) induced a rapid activation of cultured mast cells. Human HMC-1.1 and LUVA mast cell lines were cultured in the presence or absence of 1 nM E2, 20 μM desloratadine (Des), or 1 μM fulvestrant. A) Positive estrogen receptor staining of HMC-1.1 cells, Scale bar = 50 μm. B) Proliferation at 72 h of HMC-1.1 cells. Box plots show min–max and all individual data points, Mann-Whitney U test, n = 6 in each group, ns = not significant. C) Tryptase-positive HMC-1.1 cells at 5, 20, and 60 min. Box plots show min–max and all individual data points, Mann-Whitney U test, n = 10 in each group at each time point, ****P < 0.0001. Representative images of cells at different time points and from the different treatment groups. Scale bar = 20 μm. D) Tryptase-positive HMC-1.1 cells at 60 min. Box plots show min–max and all individual data points, Kruskal Wallis with Dunn’s post hoc test, n = 10 in each group, **P < 0.01, ****P < 0.0001. Representative images of the cells. Scale bar = 20 μm, E) Proteins released in the cell culture media from HMC-1.1 cells. Box plots show min–max and all individual data points, Kruskal Wallis with Dunn’s post hoc test, n = 5 in each group, *P < 0.05, **P < 0.01. F) Positive estrogen receptor staining of LUVA cells; scale bar = 50 μm. G) Proliferation of LUVA cells after 72 h. Box plots show min–max and all individual data points, Kruskal Wallis with Dunn’s post hoc test, n = 6 in each group, ***P < 0.001, ****P < 0.0001. H) Tryptase-positive LUVA cells at 5, 20, and 60 min. Box plots show min–max and all individual data points, Mann-Whitney test, n = 10 in each group at each time point, ns = not significant, **P < 0.01, ***P < 0.001 with representative images at different time points. Scale bar = 20 μm. I) Tryptase-positive LUVA cells at 60 min. Box plots show min–max and all individual data points, Kruskal Wallis with Dunn’s post hoc test; n = 10 in each group, ***P < 0.001. Scale bar = 20 μm. J) Proteins released into the cell culture media from LUVA cells. Box plots show min–max and all individual data points, Kruskal Wallis with Dunn’s post hoc test, n = 5 in each group, *P < 0.05, **P < 0.01. Next, we investigated another human mast cell line, LUVA, which is an immortalized cell line originating from CD34+ mononuclear cells of a donor with aspirin-exacerbated respiratory disease and wild-type KIT receptor expression. 33 In Figure 5F, ERα expression of LUVA cells is shown. In contrast to HMC-1.1 cells, E2 significantly increased the proliferation rate of LUVA cells, an effect counteracted by the antiestrogen fulvestrant (Figure 5G). Similarly to HMC-1.1 cells, E2 induced a rapid loss of intracellular tryptase immunoreactivity (Figure 5H), which was counteracted by desloratadine (Figure 5I). After 20 min, E2 exposure resulted in significantly increased VEGF levels, which were reversed by desloratadine (Figure 5J). However, after 60 minutes no changes were detected among the different treatments. All other proteins were below the detection limit in the LUVA cells. The cells, in both cell lines, remained present and morphologically intact during the experiments. The kinetics of E2-induced mediator secretion differed between HMC-1.1 and LUVA cells, which may reflect phenotypic differences between these two distinct human mast cell lines. Thus, in two different mast cell lines, E2 induced a time-dependent loss of intracellular tryptase immunoreactivity, accompanied by increased secretion of CCL2 and/or VEGF. Together, these findings are consistent with rapid mast cell activation and degranulation in response to E2. The role of E2 and mast cells in murine and zebrafish models of breast cancer To investigate whether E2 influences mast cell infiltration in breast cancer, we analyzed tumors grown in the presence or absence of E2 in an immunocompetent murine PyMT model. As shown in Figure 6A, E2 exposure was associated with a significant increase in the number of intratumoral mast cells. Estradiol (E2) increased mast cell number in murine breast cancer and mast cells increased ER+ breast cancer growth in zebrafish. Estrogen-dependent breast cancers were established in the mammary fat pad of oophorectomized mice. The number of mast cells was quantified (mean of 10 high-power fields, X400, in five different tumors in each group). A) PyMT tumors grown in immunocompetent FVB/N mice supplemented with physiological levels of E2 or placebo. Box plots show min-max and all individual data points, Mann-Whitney U-test, **P < 0.01, representative sections from each group; scale bar = 20 μm. B) Human ER+ MCF-7 breast cancer tumors were grown in nude mice supplemented with physiological levels of E2. In one group, pure antiestrogen fulvestrant (Fulv) was added to the E2 treatment. Box plots show min-max and all individual data points. Mann-Whitney U-test, **P < 0.01, representative sections from each group; scale bar = 20 μm. C) Transgenic zebrafish embryos were injected with MCF-7 ± LUVA mast cells ± E2 ± desloratadine (Des) in the perivitelline space, as described in Materials and Methods. Tumor growth was calculated by measuring the tumor area on day 3 compared to the tumor area on day 0 of the same embryo, n = 23-33 in each group. Box plots show min-max and all individual data points, Kruskal Wallis test followed by Dunn’s multiple comparisons test, *P < 0.05, **P < 0.01, *** P < 0.001 representative sections from each group, scale bar = 100 μm. To examine whether these results were similar in human breast cancers and whether antiestrogen treatment modulates this response, we evaluated human breast cancer xenografts established in nude mice treated with E2 alone or with E2 in combination with pure antiestrogen fulvestrant. Tumors from mice that received fulvestrant treatment exhibited significantly fewer mast cells than those exposed to E2 alone (Figure 6B). Although nude mice lack functional T cells, innate immune populations, including mast cells, are preserved in this model. Next, we investigated the functional contribution of mast cells to primary tumor growth using a zebrafish xenograft model. Human ER+ MCF-7 breast cancer cells were implanted alone or in combination with human LUVA mast cells into the perivitelline space of 2-day-old zebrafish embryos, followed by treatment with E2 and/or desloratadine. Tumor growth was quantified three days after implantation. As shown in Figure 6C, co-implantation of LUVA mast cells enhanced tumor growth compared to cancer cells alone, indicating a tumor-promoting effect of mast cells per se in this model. This effect was further potentiated by E2, which is consistent with the estrogen-dependent activation of mast cell pro-tumorigenic functions. Importantly, desloratadine abolished the E2 and mast cell–mediated enhancement of tumor growth. Increased numbers of mast cells and levels of extracellular SCF in ER+ human breast cancer Next, we assessed the clinical relevance of mast cells in human breast cancers. ER+ tumors were analyzed for mast cell density within the tumor and in matched adjacent normal breast tissues. As shown in Figure 7A, ER+ breast cancers exhibited a significantly higher number of mast cells and significantly elevated SCF levels than adjacent normal breast tissues (Figure 7B). Human ER + breast cancer exhibited increased mast cell numbers and in vivo levels of SCF. A) Tryptase-positive mast cells in human ER+ breast cancer compared with normal adjacent breast tissue (n = 12, Wilcoxon paired test, ***P < 0.001, representative sections from each group, scale bar = 50 μm). B) Extracellular proteins in breast cancers and adjacent normal tissues were sampled in situ by microdialysis in 12 postmenopausal patients with ER+ breast cancer. SCF (KITL) was quantified as described in the Materials and Methods (n = 12, Wilcoxon paired test, **P < 0.01). C) Increased expression of SCF (KITL) was associated with worse overall survival of patients with ER+ breast cancer (Kaplan-Meier Plotter database). To further evaluate the clinical significance of SCF expression for the survival of patients with breast cancer, we used the publicly available Kaplan–Meier Plotter database. These results support the importance of SCF expression, as increased levels were associated with reduced overall survival in patients with ER+ breast cancer (Figure 7C). Summary In Figure 8 our results are summarized in a graphical abstract. Graphical summary of the estrogen-associated breast tissue microenvironment and mast cell activation. Schematic summary of the main findings of the study. In vivo microdialysis and proteomic profiling of human breast tissue identified an E2-associated extracellular protein signature enriched for proteins related to mast cell biology. Complementary ex vivo and in vitro experiments demonstrated E2-induced increases in mast cell number and activity, while murine breast cancer models showed estrogen-dependent modulation of intratumoral mast cell density. In the zebrafish xenograft model, mast cells promoted ER+breast cancer growth, which was further enhanced by E2 and inhibited by desloratadine. Human ER+breast cancers exhibited increased mast cell density and extracellular SCF levels compared with adjacent normal breast tissue. Together, the findings support mast cells as estrogen-responsive components of the breast tissue microenvironment. Created in BioRender. Dabrosin, C. (2026) https://BioRender.com/8aoug2t. Discussion Here, we identified a tissue-specific protein signature consistent with estrogen-dependent regulation of mast cells in the normal human breast microenvironment using in situ sampling of extracellular proteins from live human breast tissue combined with targeted proteomic profiling. These in vivo observations were supported by ex vivo culture of intact normal human breast tissue, in which E2 exposure increased mast cell numbers and was associated with reduced intracellular tryptase immunoreactivity and altered cytokine secretion. In vitro, E2 induced rapid loss of intracellular tryptase immunoreactivity accompanied by increased secretion of inflammatory cytokines in both wild-type c-KIT LUVA mast cells and c-KIT-mutated HMC-1.1 mast cells. These effects were attenuated by the antihistamine desloratadine. Additionally, E2 increased the proliferation of LUVA cells. In parallel, in murine models of ER+ breast cancer, E2 increased mast cell accumulation in tumor tissue in immunocompetent mice, whereas fulvestrant added to the E2 treatment reduced mast cell numbers in tumor tissue of immunocompromised mice. Functionally, mast cells promoted primary tumor growth in a zebrafish xenograft model, which was further augmented by E2 and abolished by desloratadine. Consistent with these experimental findings, human ER+ breast cancers exhibited elevated SCF levels and increased mast cell infiltration compared to adjacent normal tissue, and high SCF expression was associated with reduced overall survival. The role of mast cells in breast cancer progression remains controversial. Mast cells have been associated with poor prognosis in ER+ disease but favorable prognosis in ER− tumors, 15 and several studies have reported a higher mast cell density in ER+ compared with ER− breast cancers. 34 , 35 Experimental studies suggest that mast cells may influence cancer stem–like features, stromal remodeling, ER expression, and tamoxifen sensitivity, thereby promoting the aggressiveness and dissemination of ER+ tumors. 36-38 In murine models, mast cells have been shown to activate estrogen receptor signaling, resulting in a shift toward a luminal subtype of breast cancer. 39 Supporting this concept, registry-based analyzes have linked the use of antihistamines, such as desloratadine and loratadine, to improved survival in patients with breast cancer. 40 , 41 Data directly addressing mast cell activity in breast cancer risk are limited; however, in patients with type 2 diabetes, who have an elevated risk of ER+ breast cancer, antihistamine use has been associated with a dose-dependent reduction in breast cancer incidence. 42 , 43 Together, these observations are consistent with the role of mast cells not only in tumor progression but potentially also in early tumorigenesis. Incomplete understanding of normal breast physiology limits the development of effective and well-tolerated preventive strategies. Although epidemiologic data clearly implicate steroid exposure in the risk of ER+ breast cancer, long-term antiestrogen therapy is restricted by adverse effects. Therefore, elucidation of steroid-dependent stromal mechanisms may provide alternative targets for prevention. The breast microenvironment is an important determinant of progression from pre-invasive to invasive disease as demonstrated by autopsy studies showing high prevalence of ductal carcinoma in situ (DCIS) in populations with low frequency of invasive disease. 7 , 44 , 45 Microdialysis enables direct sampling of soluble extracellular proteins mirroring the inter-cellular crosstalk in a preserved local live tissue microenvironment. Sampling of extracellular proteins from both breast tissue and abdominal subcutaneous fat in the same individual allowed us to distinguish tissue-specific alterations from systemic alterations. 22 , 46 Proteomic profiling of microdialysis samples revealed that estrogen exposure in breast tissue predominantly altered pathways related to cytokine signaling and leukocyte chemotaxis. After excluding proteins that were similarly regulated in subcutaneous fat, 202 breast-specific proteins remained, with persistent enrichment of cytokine-mediated pathways. Co-expression analysis demonstrated coordinated regulation of proteins involved in cytokine–receptor interactions and cancer-associated pathways, including SCF and members of the TNF receptor superfamily. Correlation analyzes further linked the local E2 levels to mast cell–associated pathways. Of the 202 breast-specific proteins, 150 were significantly correlated with local E2 concentrations and 56 of these were associated with mast cell function. The strongest correlations were observed for CD63, NDRG1, and SCF, which are central to mast cell biology. 47 , 48 CD63 regulates degranulation and leukocyte recruitment, 49-51 NDRG1 contributes to mast cell maturation and activation, 52 and SCF regulates mast cell proliferation, differentiation, and activation. 53 Additional E2-correlated proteins included IL-10, JUN, LILRB4, TNFα, CCL2, CCL4, CCL23, and VEGF, which are mediators linked to mast cell activation, inflammation, and tumor-promoting processes. 54-58 Collectively, these in situ data indicate that estrogen exposure is closely associated with a coordinated mast cell–related signaling network in the human breast. It may be argued that the marked separation in E2 concentrations between the groups would predictably result in significant correlations between E2 and all proteins. However, approximately one quarter of the proteins that differed significantly between pre- and postmenopausal women did not show a significant association with E2 concentrations. This demonstrates that the E2 correlations are not an inevitable consequence of the dichotomous cohort structure. Importantly, the in situ correlative findings in women have been substantiated by functional studies across complementary model systems. Ex vivo culture of intact normal human breast tissue demonstrated that E2 increased the number of mast cells and enhanced the secretion of CCL2, CCL4, CCL23, TNFα, and VEGF. In vitro, E2 induced a rapid loss of intracellular tryptase immunoreactivity accompanied by increased secretion of inflammatory mediators in two ER-expressing mast cell lines, findings consistent with rapid mast cell activation and degranulation. The rapid onset of this response is consistent with non-genomic estrogen signaling, as previously described in mast cells. 59 , 60 Notably, E2-induced mast cell degranulation was abrogated by the antihistamine desloratadine in vitro. In parallel, E2 increased the proliferation of LUVA cells, whereas no proliferative response was observed in HMC-1 cells. This differential response may reflect differences in ER signaling between the two mast cell lines. However, the effects in LUVA cells suggest that estrogen can regulate mast cell activity through both rapid and long-term mechanisms. These findings were further confirmed using in vivo model systems of breast cancer. In the immunocompetent murine breast cancer model, E2 increased mast cell accumulation within tumor tissue, whereas in the immunocompromised MCF-7 model, fulvestrant reduced mast cell accumulation in E2-treated tumors. In a zebrafish xenograft model, mast cells enhanced ER+tumor growth, which was further potentiated by E2 and effectively suppressed by desloratadine. Consistent with these experimental data, human ER+ breast cancers exhibited elevated SCF levels together with increased mast cell infiltration, and high SCF expression was associated with reduced overall survival. A major strength of this study is the direct in vivo sampling of the extracellular microenvironment of live human breast cancer and normal human breast tissue, with simultaneous sampling of breast and abdominal subcutaneous tissue enabling distinction between breast-specific and systemic alterations. Moreover, the human in situ findings were investigated across complementary ex vivo, in vitro, and in vivo experimental models. Several limitations should also be acknowledged. In the human cohort, local E2 concentrations were associated with menopausal status, and effects of E2 cannot therefore be fully separated from age-related differences. However, not all proteins differing between pre- and postmenopausal women correlated with local breast E2. Human mast cell lines cannot fully recapitulate the phenotype of mast cells residing in human breast tissue, and their differential responses to E2 may reflect differences in their cellular phenotypes and ER signaling. The human studies were conducted at a single center in relatively modest cohorts, and the zebrafish xenograft represents a short-term experimental model of tumor growth. Finally, the concentration of desloratadine used in vitro should not be directly extrapolated to clinically achievable systemic exposure, and further studies are required to define the concentration-response relationship and translational relevance of mast cell inhibition in vivo. Conclusions Together, these findings identify mast cells as estrogen-responsive stromal mediators and support the exploration of mast cell–dependent pathways as a potential strategy for the prevention and treatment of estrogen-dependent breast cancer. Supplementary Material Supp Table 2 proteins related to mast cell function1.docx SuppTable 1 proteins tissue specific.xlsx SuppTable 1 proteins tissue specific.xlsx