ZNF831 suppresses triple-negative breast cancer progression through NLRP3-associated pyroptotic signaling and M1-like macrophage phenotypic remodeling Authors: Qingqing Liu, Jun Fan, Wantao Peng, Shuangwu Deng, Yupei Tan, Fei Guo, Jiangxue Zhong, Shimeng Zhou, Wenbin Liu, Yan Xu Published: 2026 ABSTRACT BackgroundTriple-negative breast cancer (TNBC) remains difficult to treat because of poor immunogenicity and an immunosuppressive tumor microenvironment (TME). Pyroptotic signaling can enhance antitumor immunity through inflammatory cytokine release, but upstream regulators linking this process to macrophage phenotypic remodeling in TNBC remain incompletely understood. We investigated whether zinc finger protein 831 (ZNF831) restrains TNBC progression in association with enhanced NLR family pyrin domain containing 3 (NLRP3)-related pyroptotic signaling and macrophage phenotypic remodeling.MethodsWe integrated The Cancer Genome Atlas (TCGA-BRCA), Gene Expression Omnibus bulk datasets (GSE103091, GSE176078), and breast cancer single-cell RNA-sequencing datasets from Tumor Immune Single-cell Hub 2 (BRCA_GSE114727_inDrop, BRCA_GSE148673, BRCA_GSE150660, and BRCA_GSE161529) to evaluate ZNF831 expression, prognosis, pathway enrichment, and macrophage-related immune features. ZNF831 was overexpressed or silenced in MDA-MB-231 and 4T1-luc cells to assess macrophage recruitment, macrophage phenotypic changes, NLRP3-associated pyroptotic signaling, and IL-1β/IL-18 release. Chromatin immunoprecipitation (ChIP)-qPCR, actinomycin D chase assays, and RNA immunoprecipitation (RIP)-qPCR were performed primarily in MDA-MB-231 cells to evaluate promoter association, mRNA stability, and transcript interaction. In vivo efficacy was assessed in 6-week-old female BALB/c mice bearing orthotopic 4T1-luc tumors treated intratumorally with MCC950 or PBS.ResultsAcross cohorts, higher ZNF831 expression was associated with favorable outcome, an M1-skewed macrophage signature, and enrichment of pyroptosis-related pathways. In vitro, ZNF831 overexpression enhanced macrophage recruitment, shifted macrophage phenotypes toward a more pro-inflammatory M1-like state while suppressing M2-like features, promoted NLRP3-associated pyroptotic signaling, and increased IL-1β and IL-18 release, whereas ZNF831 knockdown produced opposite trends. Mechanistically, ZNF831 showed association with the NLRP3 promoter and selective enrichment of NLRP3 mRNA, consistent with multi-level regulation of NLRP3 expression. In vivo, pharmacologic inhibition with MCC950 attenuated the antitumor effects and macrophage-related changes associated with ZNF831 overexpression.ConclusionsThese findings support a role for the ZNF831–NLRP3 axis in shaping a more inflammatory TNBC microenvironment and suggest that ZNF831 may contribute to macrophage M1-like phenotypic remodeling through NLRP3-associated pyroptotic signaling.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13058-026-02307-7. Background Triple-negative breast cancer (TNBC) is among the most aggressive subtypes of breast cancer. It is defined by the absence of estrogen receptor (ER) and progesterone receptor (PR) expression together with the absence of HER2 overexpression/amplification rather than a complete lack of HER2 protein expression [1, 2]. This molecular profile limits the use of endocrine therapy and HER2-targeted treatment, and contributes to its high invasive potential, early metastatic tendency, and poor clinical outcome [1]. Although the combination of immune checkpoint blockade and chemotherapy has improved outcomes in a subset of patients with programmed death-ligand 1 (PD-L1)-positive TNBC, a substantial proportion of patients still derive limited benefit because of poorly inflamed or immune-excluded tumor microenvironments [3–5]. These clinical challenges underscore the need to identify additional regulators of antitumor immunity in TNBC. The tumor microenvironment (TME) plays a central role in TNBC progression and therapeutic resistance [5]. Among immune populations in the TME, tumor-associated macrophages (TAMs) are among the most abundant myeloid cell populations and have been implicated in multiple aspects of tumor biology, including inflammation, angiogenesis, invasion, metastasis, and treatment response [5–8]. The classical M1/M2 framework provides a useful conceptual model for describing macrophage states: classically activated M1-like macrophages are generally associated with pro-inflammatory and antitumor functions, whereas alternatively activated M2-like macrophages are often linked to tissue remodeling and tumor-promoting activities [8, 9]. However, macrophage phenotypes in vivo exist along a dynamic continuum rather than as two rigidly separated states [10, 11]. In TNBC, although several studies have associated M1-like macrophage features with more favorable outcomes, this relationship is not uniformly consistent across cohorts [12]. Therefore, identifying tumor-intrinsic regulators associated with a more pro-inflammatory macrophage phenotype remains an important question. Pyroptosis is a lytic form of programmed cell death mediated by gasdermin family proteins and characterized by cell swelling, membrane rupture, and inflammatory cytokine release [13]. A central pathway underlying canonical pyroptosis involves activation of the NLR family pyrin domain containing 3 (NLRP3) inflammasome. Upon sensing danger-associated or pathogen-associated signals, NLRP3 recruits apoptosis-associated speck-like protein (ASC) and caspase-1, leading to caspase-1 activation, cleavage of gasdermin D (GSDMD), pore formation in the plasma membrane, and maturation and release of interleukin-1β (IL-1β) and interleukin-18 (IL-18) [13, 14]. Through these inflammatory outputs, pyroptotic signaling has the potential to reshape the immune contexture of the TME and influence antitumor immunity [14]. However, the upstream regulators that connect pyroptotic signaling with macrophage phenotypic remodeling in TNBC remain incompletely understood. Zinc finger protein 831 (ZNF831) is a KRAB domain-containing zinc finger transcription factor. Previous studies have shown that ZNF831 is downregulated in breast cancer and can suppress malignant phenotypes while promoting apoptosis [15]. In TNBC immunogenomic analyses, ZNF831 has also been associated with immune-active tumor states and improved prognosis [16]. These observations raise the possibility that ZNF831 may participate in immune regulation beyond its previously described tumor-intrinsic effects. In the present study, we integrated bulk transcriptomic and single-cell datasets with in vitro and in vivo experiments to investigate whether ZNF831 is associated with NLRP3-related pyroptotic signaling and macrophage phenotypic remodeling in TNBC. Together, these observations suggest that ZNF831 may help shape a more inflammatory TNBC microenvironment through NLRP3-associated pyroptotic signaling. Materials and methods Cell lines and culture Human TNBC cell line MDA-MB-231 (Procell Life Science & Technology Co., Ltd., Cat. No. STCC10503, RRID: CVCL_0062) was cultured in Leibovitz’s L-15 Medium (Gibco, Cat. No. 11415064) supplemented with 10% fetal bovine serum (FBS; Bio-channel, Cat. No. BC-SE-FBS08) and 1% penicillin-streptomycin (Dowobio Cat. No. DW0022) at 37 °C in a CO₂-free atmosphere. Mouse mammary cancer cell line 4T1-luc (Procell Life Science & Technology Co., Ltd., Cat. No. STCC20022, RRID: CVCL_J239) was cultured in RPMI-1640 Medium (Gibco, Cat. No. 11875093) with the same supplements at 37 °C in a humidified atmosphere with 5% CO₂. Mouse macrophage cell line RAW264.7 (Procell Life Science & Technology Co., Ltd., Cat. No. STCC20020P, RRID: CVCL_0493) was cultured in Dulbecco’s Modified Eagle Medium (Gibco, Cat. No. 11995065) with the same supplements at 37 °C in a humidified atmosphere with 5% CO₂. All cells were authenticated by short tandem repeat profiling. Plasmids and stable cell lines ZNF831 overexpression (pcDNA3.1-3×FLAG-ZNF831-neo), empty vector, and shRNA plasmids (pLKO.1-puro) were constructed in-house (sequences provided in Table S1). Transient transfection was performed using NanoTrans™ Transfection Reagent 3000 (CYTOCH, Cat. No. CT0006). Stable cells were selected with G418 (500 µg/mL for MDA-MB-231; 400 µg/mL for 4T1-luc). Overexpression and knockdown efficiency were verified by reverse transcription quantitative polymerase chain reaction (RT-qPCR). Transwell migration assay For migration assays, 5 × 104 tumor cells (MDA-MB-231 or 4T1-luc) were seeded in the lower chamber, and 2 × 105 RAW264.7 cells were added to the upper insert (Corning Transwell, 6.5 mm diameter, 8 μm pore size polycarbonate membrane; Cat. No. 3422). After 24 h, migrated RAW264.7 cells were fixed, stained with crystal violet, photographed under a light microscope, and quantified using ImageJ (RRID: SCR_003070) software. Macrophage phenotyping after conditioned medium stimulation RAW264.7 cells were stimulated for 24 h with 50% tumor cell-conditioned medium. Macrophage phenotypes were evaluated by flow cytometry, RT-qPCR, and western blotting. For flow cytometric analysis, RAW264.7 cells were stained with PE-conjugated anti-mouse F4/80 together with FITC-conjugated anti-mouse CD86 for the M1-like readout, or with PE-conjugated anti-mouse F4/80 together with APC-conjugated anti-mouse CD206 for the M2-like readout. Flow cytometry antibodies are listed in Table S3. Debris was excluded by FSC/SSC gating, followed by singlet selection, and the percentages of F4/80⁺CD86⁺ cells and F4/80⁺CD206⁺ cells were quantified. Unstained and single-stained controls were used to define quadrant gates, which were applied consistently across groups within each experiment. Representative gating strategies are shown in Fig. S2. For M1-associated marker analysis, CD86, CD80, iNOS, and IL-6 were assessed by RT-qPCR, and CD86 and iNOS were assessed by western blotting. For M2-associated marker analysis, CD206, ARG1, CD163, and IL-10 were assessed by RT-qPCR, and CD206 and ARG1 were assessed by western blotting. Primer sequences are listed in Table S2, and antibodies are listed in Table S3. Molecular analyses Total RNA was isolated using TRIzol Reagent (Invitrogen, Cat. No. 15596026), reverse-transcribed into cDNA using the ABScript™ Neo RT Master Mix for qPCR with gDNA Remover (Abclonal, Cat. No. RK20433), and quantified by RT-qPCR using 2X Universal SYBR Green Fast qPCR Mix (Abclonal, Cat. No. RK21203) with β-actin as the reference gene (primers listed in Table S2). Proteins were lysed in RIPA buffer (Beyotime, Cat. No. P0013D) supplemented with protease and phosphatase inhibitors (Beyotime, Cat. No. P1045). Lysates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Cat. No. IPVH00010), and detected using enhanced chemiluminescence (ECL; Thermo Fisher Scientific, Cat. No. 32106). Primary and secondary antibodies are listed in Table S3. Extracellular mature IL-1β and IL-18 in concentrated culture supernatants were also assessed by western blotting. Secreted IL-1β and IL-18 levels in cell culture supernatants were measured using ELISA kits: Human IL-1β ELISA Kit (Huabio, Cat. No. EH0006), Human IL-18 ELISA Kit (Huabio, Cat. No. EH0053), Mouse IL-1β ELISA Kit (Huabio, Cat. No. EM0029), and Mouse IL-18 ELISA Kit (Huabio, Cat. No. EM0034). LDH release assay LDH release was measured using the LDH Cytotoxicity Assay Kit (Beyotime, Shanghai, China; Cat. No. C0017) according to the manufacturer’s instructions. Cells were seeded in 96-well plates and cultured under the indicated conditions. At the indicated time points, culture supernatants were collected and transferred to a fresh 96-well plate for LDH detection. Absorbance was measured at 490 nm with a reference wavelength above 600 nm for background correction. A background blank control, sample control, and maximum enzyme activity control were included in each experiment, and LDH cytotoxicity (%) was calculated according to the manufacturer’s formula. LDH cytotoxicity was used as an indicator of lytic cell death. For overexpression and knockdown experiments, LDH release was assessed in MDA-MB-231 and 4T1-luc cells over time. For MCC950 experiments, LDH release was measured in Vector-, ZNF831-, and ZNF831 + MCC950-treated cells at the indicated time points. mRNA stability assay For mRNA stability assays, MDA-MB-231 cells were treated with 5 µg/mL actinomycin D (Sigma-Aldrich, Cat. No. A9415) to inhibit transcription, and total RNA was collected at 0, 2, 4, 6, and 8 h after treatment. NLRP3 mRNA levels were quantified by RT-qPCR and normalized to the corresponding baseline level at 0 h. The decay profile was used to assess the apparent stability of NLRP3 mRNA in vector- and ZNF831-overexpressing cells. Chromatin immunoprecipitation (ChIP) assayc ChIP was performed using the SimpleChIP® Enzymatic Chromatin IP Kit (Cell Signaling Technology) according to the manufacturer’s instructions. Briefly, FLAG-ZNF831-overexpressing MDA-MB-231 cells were cross-linked, chromatin was digested and sheared, and immunoprecipitation was performed with anti-FLAG antibody or IgG control. qPCR was then carried out using primers targeting three predicted regions within the human NLRP3 promoter as well as the RPL30 control locus. ChIP signals were expressed as percentage of input (% input). Histone H3 was used as a chromatin recovery control. Corresponding HEK293T ChIP assays were performed as supplementary technical validation and are shown in Fig. S4C–F. RNA Immunoprecipitation (RIP) assay RIP was performed under RNase-free conditions in FLAG-ZNF831-overexpressing MDA-MB-231 cells. Cell lysates were incubated with anti-FLAG antibody or IgG control, and RNA recovered from immunoprecipitates was purified for RT-qPCR analysis. Enrichment of NLRP3 mRNA was assessed in the FLAG pull-down, while GAPDH mRNA was used as a negative control transcript. Corresponding HEK293T RIP assays were performed as supplementary technical validation and are shown in Figure S4G, H. In vivo studies All procedures were approved by the Institutional Animal Care and Use Committee of Army Medical University. 4T1-luc cells (5 × 10⁵) were orthotopically transplanted into the fourth mammary fat pad of 6-week-old female BALB/c mice (n = 5/group). MCC950 (10 mg/kg) or PBS was given intratumorally every other day. Tumor volume (measured as (length × width²)/2) was monitored every 3 days. Mice were killed on day 21 for tumor weight and subsequent analysis. Tumor-associated bioluminescence was assessed using an IVIS Spectrum system (PerkinElmer) after intraperitoneal injection of D-luciferin (150 mg/kg; Gold Biotechnology, Cat. No. LUCK-1G). For quantitative analysis, regions of interest (ROIs) were defined over the tumor-bearing mammary fat pad region on each bioluminescence image using the same ROI placement strategy across groups. The bioluminescence signal within each ROI was quantified as counts per second (CPS). For Fig. 7D, CPS values were normalized to the mean value of the Vector group and presented as relative tumor-associated BLI signal. Histology Resected tumor tissues were fixed in 10% neutral-buffered formalin, embedded in paraffin, and sectioned at 5 μm thickness. Hematoxylin and eosin (H&E) staining was performed for routine histopathological examination. For immunohistochemistry (IHC) and immunofluorescence (IF), tumor sections were subjected to antigen retrieval in citrate buffer (pH 6.0), incubated overnight at 4 °C with primary antibodies against F4/80, CD86 and/or CD206, and then detected using Servicebio’s proprietary HRP-conjugated secondary antibody kits and DAB substrate system (IHC) or Alexa Fluor 488/594-conjugated secondary antibodies (IF). Images were acquired using a fluorescence microscope and quantified using ImageJ software by measuring the percentage of positive staining area or the number of positive cells in at least five random fields per section. Bioinformatics Publicly available datasets were used for analysis of macrophage-related immune features in breast invasive carcinoma (BRCA). Single-cell RNA-seq datasets from Tumor Immune Single-cell Hub 2 (BRCA_GSE114727_inDrop, BRCA_GSE148673, BRCA_GSE150660, and BRCA_GSE161529) were used to evaluate overall immune-cell composition and macrophage proportions. The GEO dataset GSE176078 was used to compare macrophage proportions across BRCA molecular subtypes. Bulk transcriptomic datasets from TCGA-BRCA and GSE103091 were used to estimate macrophage-related immune features and to screen correlated genes in TNBC. Survival analyses were performed using the Kaplan-Meier Plotter (KM-Plotter) online tool. For analyses requiring stratification by ZNF831 expression, including survival analysis, immune-feature comparison, and pyroptosis score comparison, TNBC samples were divided into ZNF831-high and ZNF831-low groups using the median ZNF831 expression value within the corresponding cohort as the cutoff. This median-based stratification approach was applied consistently throughout the manuscript unless otherwise specified. Pathway enrichment and correlation analyses were performed in the TNBC subset of TCGA-BRCA. Code is available upon request. Statistics All experiments were performed with at least three independent biological replicates unless otherwise stated. Data are presented as mean ± SD. Comparisons between two groups were performed using two-tailed unpaired Student’s t-test, whereas comparisons among three or more groups were performed using one-way ANOVA followed by Tukey’s multiple-comparisons test, as appropriate. LDH release assays were analyzed separately at each time point by one-way ANOVA with Tukey’s multiple-comparisons test. Tumor growth curves were analyzed using repeated-measures two-way ANOVA. Assumptions for parametric testing were assessed where applicable before test selection. Statistical analyses were performed using GraphPad Prism (RRID: SCR_002798), and a two-sided p value < 0.05 was considered statistically significant. Results Integrative immune-feature screening nominates ZNF831 as a TNBC-associated candidate linked to favorable outcome To characterize the breast cancer immune microenvironment and identify candidate regulators, we first integrated multiple publicly available scRNA-seq datasets (GSE114727_inDrop, GSE148673, GSE150660, and GSE161529). Across these datasets, macrophages consistently ranked among the most abundant or second most abundant immune populations, comparable to or only exceeded by T cells (Figs. 1A and S1A), highlighting their potential importance in the breast cancer microenvironment. To further examine macrophage composition across molecular subtypes, we analyzed the GSE176078 dataset with annotated subtype information. The cellular composition differed markedly among subtypes, and macrophages accounted for 15.8% of total cells in TNBC, compared with 8.9% in HER2 + tumors and 10.0% in ER+ tumors (Fig. 1B). The corresponding cell-type annotation framework for the GSE176078 dataset is shown in Fig. S1B. Given the known plasticity of tumor-associated macrophages and the reported association of M1-like macrophage features with more favorable outcomes in several solid tumors, including TNBC [12], we hypothesized that upstream regulators associated with M1-related rather than M2-related macrophage features might represent biologically relevant candidates. To identify such factors, we analyzed TNBC samples from two independent cohorts (TCGA-BRCA and GSE103091) and screened for genes that simultaneously satisfied the following criteria: (i) significant positive correlation with classical M1-related markers (iNOS, CXCL10, TNF, etc.), (ii) significant negative correlation with M2-related markers (ARG1, MRC1, IL10, etc.), and (iii) association with significantly improved survival in TNBC (hazard ratio [HR] 95% confidence interval [CI] excluding 1). The intersection of the two screens yielded nine candidate genes (Fig. 1C–E). To further assess subtype specificity, we performed subtype-stratified survival analyses using the Kaplan–Meier Plotter (KM-Plotter) database. Among the candidates, ZNF831 showed the strongest association with prolonged overall survival in TNBC (HR = 0.42, 95% CI 0.30–0.58, P = 4.4 × 10−8), while showing limited or no prognostic value in Luminal A, HER2-positive, or other subtypes (Figs. 1F and S1C). These findings nominate ZNF831 as a candidate TNBC-associated immune regulator. To further validate the relationship between ZNF831 and macrophage-related immune features, we performed immune deconvolution on TNBC subsets of TCGA-BRCA and GSE103091 using CIBERSORT. ZNF831 mRNA levels showed a strong positive association with M1-related macrophage features and a negative association with M2-related features (Fig. 1G, H). As shown in Fig. 1I, the correlation coefficient with the M1-like macrophage signature (|r| = 0.51) was greater than that with the M2-related signature (|r| = 0.35). Collectively, these multi-layered bioinformatic analyses identify ZNF831 as a TNBC-relevant candidate associated with an M1-skewed macrophage signature and improved clinical outcome. Fig. 1High ZNF831 expression in TNBC is associated with an M1-skewed macrophage signature and favorable prognosis. A Immune-cell composition in breast cancer from integrated scRNA-seq datasets (GSE114727_inDrop, GSE148673, GSE150660, GSE161529), showing a high proportion of macrophages. B Analysis of the GSE176078 dataset showing increased macrophage proportions in TNBC relative to other breast cancer subtypes. C, D Correlation analysis in TNBC samples from TCGA-BRCA and GSE103091 identifying genes positively associated with M1-related signatures and negatively associated with M2-related signatures. E Venn diagram showing candidate genes identified by intersection analysis and prognostic filtering. F Kaplan-Meier survival analysis in TNBC stratified by median ZNF831 expression. G, H Comparison of macrophage-related immune features between ZNF831-high and ZNF831-low TNBC groups defined using the median ZNF831 expression cutoff. I Violin plots comparing macrophage-related signatures between ZNF831-high and ZNF831-low groups defined by the median ZNF831 expression cutoff High ZNF831 expression in TNBC is associated with an M1-skewed macrophage signature and favorable prognosis. A Immune-cell composition in breast cancer from integrated scRNA-seq datasets (GSE114727_inDrop, GSE148673, GSE150660, GSE161529), showing a high proportion of macrophages. B Analysis of the GSE176078 dataset showing increased macrophage proportions in TNBC relative to other breast cancer subtypes. C, D Correlation analysis in TNBC samples from TCGA-BRCA and GSE103091 identifying genes positively associated with M1-related signatures and negatively associated with M2-related signatures. E Venn diagram showing candidate genes identified by intersection analysis and prognostic filtering. F Kaplan-Meier survival analysis in TNBC stratified by median ZNF831 expression. G, H Comparison of macrophage-related immune features between ZNF831-high and ZNF831-low TNBC groups defined using the median ZNF831 expression cutoff. I Violin plots comparing macrophage-related signatures between ZNF831-high and ZNF831-low groups defined by the median ZNF831 expression cutoff ZNF831-modulated TNBC cell secretomes favor macrophage recruitment and a pro-inflammatory phenotypic shift in vitro To determine whether ZNF831 in TNBC cells regulates macrophage recruitment, we first performed transwell migration assays (Fig. 2A). MDA-MB-231 or 4T1-luc cells transfected with pcDNA3.1-3×FLAG-ZNF831 (ZNF831 group) or empty vector (Vector group) were seeded in the lower chamber 24 h post-transfection, and RAW264.7 macrophages were added to the upper chamber. After 24 h of co-culture, ZNF831 overexpression in tumor cells markedly increased RAW264.7 transwell migration in both cell lines compared with vector control (Fig. 2B). Conversely, stable knockdown of ZNF831 using two distinct shRNAs (shZNF831 #1 and shZNF831 #2) significantly reduced the number of migratory macrophages compared with the shNC group (Fig. 2C). These results indicate that ZNF831 contributes to macrophage recruitment by TNBC cells. We next examined whether ZNF831 also affects macrophage phenotypes. RAW264.7 cells were stimulated for 24 h with conditioned medium (CM) derived from the indicated TNBC cell groups (workflow in Fig. 2D). Flow cytometry showed that CM from ZNF831-overexpressing tumor cells increased the proportion of F4/80⁺CD86⁺ macrophages, whereas the opposite trend was observed in the ZNF831-knockdown groups (Fig. 2E). Consistently, RT-qPCR demonstrated increased expression of M1-associated genes, including CD86, CD80, iNOS, and IL-6, in macrophages exposed to ZNF831-overexpressing CM, whereas reciprocal changes were observed after ZNF831 knockdown (Fig. 2F, G). Western blotting likewise showed increased expression of CD86 and iNOS in the overexpression setting and decreased expression in the knockdown groups (Fig. 2H, I). We also assessed M2-like features to provide a more balanced phenotypic characterization. Flow cytometry showed a reduction in the F4/80⁺CD206⁺ population in macrophages treated with CM from ZNF831-overexpressing tumor cells, whereas ZNF831 knockdown showed the opposite trend (Fig. 2J). Similarly, M2-associated markers, including CD206, ARG1, CD163, and IL-10, were reduced at the mRNA level, and CD206 and ARG1 were decreased at the protein level in macrophages exposed to ZNF831-overexpressing CM (Fig. 2K–N). Together, these findings indicate that ZNF831-modulated TNBC cells promote macrophage recruitment and shift macrophage phenotypes toward a more pro-inflammatory, M1-like state while suppressing M2-like features. Fig. 2ZNF831 promotes macrophage recruitment and shifts macrophage phenotypes toward an M1-like state in vitro. A Schematic diagram of the transwell migration assay. B, C Transwell migration assays of RAW264.7 macrophages toward MDA-MB-231 or 4T1-luc cells with ZNF831 overexpression or knockdown. D Workflow for macrophage polarization using tumor cell-conditioned media. E Flow cytometry analysis of F4/80⁺CD86⁺ macrophages. F, G RT-qPCR analysis of M1-associated genes (CD86, CD80, iNOS, and IL-6). H, I Western blot analysis of M1-associated proteins iNOS and CD86. J Flow cytometry analysis of F4/80⁺CD206⁺ macrophages. K, L RT-qPCR analysis of M2-associated genes (CD206, ARG1, CD163, and IL-10). M, N Western blot analysis of M2-associated proteins CD206 and ARG1. Data are presented as mean ± SD from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 ZNF831 promotes macrophage recruitment and shifts macrophage phenotypes toward an M1-like state in vitro. A Schematic diagram of the transwell migration assay. B, C Transwell migration assays of RAW264.7 macrophages toward MDA-MB-231 or 4T1-luc cells with ZNF831 overexpression or knockdown. D Workflow for macrophage polarization using tumor cell-conditioned media. E Flow cytometry analysis of F4/80⁺CD86⁺ macrophages. F, G RT-qPCR analysis of M1-associated genes (CD86, CD80, iNOS, and IL-6). H, I Western blot analysis of M1-associated proteins iNOS and CD86. J Flow cytometry analysis of F4/80⁺CD206⁺ macrophages. K, L RT-qPCR analysis of M2-associated genes (CD206, ARG1, CD163, and IL-10). M, N Western blot analysis of M2-associated proteins CD206 and ARG1. Data are presented as mean ± SD from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 Transcriptomic enrichment points to pyroptosis-related inflammatory signaling as the principal pathway associated with ZNF831 in TNBC To identify molecular pathways through which ZNF831 may influence the TNBC immune microenvironment, we performed pathway enrichment analysis in TNBC samples from TCGA-BRCA stratified by median ZNF831 expression. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed that the NOD-like receptor signaling pathway was the most significantly enriched pathway in ZNF831-high tumors, followed by NF-κB, TNF, and IL-17 signaling (Fig. 3A). Consistently, gene set variation analysis (GSVA) using Molecular Signatures Database (MSigDB) Hallmark gene sets identified inflammatory response and TNFα/NF-κB signaling among the top pathways positively associated with ZNF831 expression (Fig. 3B). Given the prominence of NOD-like receptor signaling, we next calculated a consensus pyroptosis activity score based on key pyroptosis-related genes, including NLRP3, AIM2, NLRC4, NLRP1, ASC, caspase-1, GSDMD, IL-1β, and IL-18, among others [11]. ZNF831-high tumors exhibited significantly higher pyroptosis scores than ZNF831-low tumors (Fig. 3C). Spearman correlation analysis of individual pyroptosis-related genes showed that the four conventional inflammasome sensors NLRP3, AIM2, NLRC4, and NLRP1 displayed the strongest positive associations with ZNF831 (r = 0.61–0.69, Fig. 3D, E). Hierarchical clustering of the full pyroptosis gene correlation matrix further indicated that ZNF831 clustered closely within a module associated with these four sensors, with NLRP3 showing the strongest pairwise correlation (r = 0.69, Fig. 3F). Collectively, these multi-layered cross-cohort bioinformatic analyses indicate that high ZNF831 expression in TNBC is strongly associated with inflammatory and pyroptosis-related signaling, particularly the NLRP3 axis. Fig. 3Bioinformatic analyses reveal a robust association between ZNF831 and pyroptosis-related signaling pathways in TNBC. A KEGG pathway enrichment analysis in ZNF831-high versus ZNF831-low TNBC samples from TCGA-BRCA. B GSVA of Hallmark pathways associated with ZNF831 expression. C Comparison of pyroptosis-related activity scores between ZNF831-high and ZNF831-low TNBC groups defined by median ZNF831 expression. D, E Spearman correlation analysis between ZNF831 and individual pyroptosis-related genes. F Hierarchical clustering of the pyroptosis-related gene correlation matrix Bioinformatic analyses reveal a robust association between ZNF831 and pyroptosis-related signaling pathways in TNBC. A KEGG pathway enrichment analysis in ZNF831-high versus ZNF831-low TNBC samples from TCGA-BRCA. B GSVA of Hallmark pathways associated with ZNF831 expression. C Comparison of pyroptosis-related activity scores between ZNF831-high and ZNF831-low TNBC groups defined by median ZNF831 expression. D, E Spearman correlation analysis between ZNF831 and individual pyroptosis-related genes. F Hierarchical clustering of the pyroptosis-related gene correlation matrix ZNF831 overexpression strengthens pyroptosis-associated execution and inflammatory cytokine release in TNBC cells Building on the bioinformatic analyses suggesting a strong association between ZNF831 and pyroptosis-related signaling, we next sought to determine whether ZNF831 promotes execution of pyroptotic cell death in TNBC cells. Phase-contrast imaging showed that ZNF831-overexpressing MDA-MB-231 and 4T1-luc cells displayed prominent cell swelling and bubble-like membrane protrusions, whereas these morphological changes were much less evident in vector control cells (Fig. 4A). Consistent with these morphological changes, LDH release assays showed that ZNF831 overexpression increased, whereas ZNF831 knockdown decreased, lytic cell death over time in both MDA-MB-231 and 4T1-luc cells (Fig. 4B, C). RT-qPCR further showed that ZNF831 overexpression increased, whereas ZNF831 knockdown decreased, the mRNA levels of NLRP3, ASC, caspase-1, GSDMD, IL-1β, and IL-18 in both cell lines (Fig. 4D, E). Western blot analysis similarly showed that ZNF831 overexpression increased, whereas ZNF831 knockdown reduced, the protein levels of NLRP3, ASC, caspase-1, cleaved caspase-1 (p20), GSDMD, and GSDMD-NT in both cell lines (Fig. 4F, G). To further assess cytokine processing and release, we examined concentrated culture supernatants by western blotting and found that ZNF831 overexpression increased, whereas ZNF831 knockdown reduced, the secretion of mature IL-1β and IL-18 in both MDA-MB-231 and 4T1-luc cells (Fig. 4H). Corresponding intracellular pro-IL-1β and pro-IL-18 protein levels are shown in Fig. S3. ELISA analysis similarly showed increased secretion of IL-1β and IL-18 in the ZNF831-overexpression groups and decreased secretion in the ZNF831-knockdown groups (Fig. 4I, J). Together, these data provide morphological, functional, and biochemical evidence that ZNF831 promotes NLRP3-associated pyroptotic signaling and lytic inflammatory responses in TNBC cells. Fig. 4ZNF831 promotes NLRP3-associated pyroptotic signaling and lytic cell death in TNBC cells. A Phase-contrast microscopy of MDA-MB-231 and 4T1-luc cells showing cell swelling and bubble-like membrane protrusions in the ZNF831-overexpression group. Scale bar, 50 μm. B, C LDH release assays showing that ZNF831 overexpression increased, whereas ZNF831 knockdown decreased, lytic cell death over time in MDA-MB-231 and 4T1-luc cells. In the overexpression panels, asterisks indicate comparisons between the ZNF831 and Vector groups at the corresponding time points. In the knockdown panels, asterisks indicate comparisons between the shZNF831 #1 and shNC groups, whereas number signs indicate comparisons between the shZNF831 #2 and shNC groups at the corresponding time points. D, E RT-qPCR analysis of NLRP3, ASC, caspase-1, GSDMD, IL-1β, and IL-18 mRNA levels in MDA-MB-231 and 4T1-luc cells with ZNF831 overexpression or knockdown. F, G Western blot analysis of NLRP3, ASC, caspase-1, cleaved caspase-1 (p20), GSDMD, and GSDMD-NT in MDA-MB-231 and 4T1-luc cells with ZNF831 overexpression or knockdown. H Western blot analysis of mature IL-1β and mature IL-18 in concentrated culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. I, J ELISA quantification of secreted IL-1β and IL-18 in culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. Corresponding intracellular pro-IL-1β and pro-IL-18 protein levels are shown in Fig. S3. Data are presented as mean ± SD from three independent experiments. For LDH assays, statistical comparisons were performed among groups at each individual time point. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001; #p < 0.05, ##p < 0.01, ###p < 0.001 ZNF831 promotes NLRP3-associated pyroptotic signaling and lytic cell death in TNBC cells. A Phase-contrast microscopy of MDA-MB-231 and 4T1-luc cells showing cell swelling and bubble-like membrane protrusions in the ZNF831-overexpression group. Scale bar, 50 μm. B, C LDH release assays showing that ZNF831 overexpression increased, whereas ZNF831 knockdown decreased, lytic cell death over time in MDA-MB-231 and 4T1-luc cells. In the overexpression panels, asterisks indicate comparisons between the ZNF831 and Vector groups at the corresponding time points. In the knockdown panels, asterisks indicate comparisons between the shZNF831 #1 and shNC groups, whereas number signs indicate comparisons between the shZNF831 #2 and shNC groups at the corresponding time points. D, E RT-qPCR analysis of NLRP3, ASC, caspase-1, GSDMD, IL-1β, and IL-18 mRNA levels in MDA-MB-231 and 4T1-luc cells with ZNF831 overexpression or knockdown. F, G Western blot analysis of NLRP3, ASC, caspase-1, cleaved caspase-1 (p20), GSDMD, and GSDMD-NT in MDA-MB-231 and 4T1-luc cells with ZNF831 overexpression or knockdown. H Western blot analysis of mature IL-1β and mature IL-18 in concentrated culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. I, J ELISA quantification of secreted IL-1β and IL-18 in culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. Corresponding intracellular pro-IL-1β and pro-IL-18 protein levels are shown in Fig. S3. Data are presented as mean ± SD from three independent experiments. For LDH assays, statistical comparisons were performed among groups at each individual time point. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001; #p < 0.05, ##p < 0.01, ###p < 0.001 Pharmacologic inhibition of NLRP3 signaling weakens the macrophage-modulating effects of ZNF831-overexpressing TNBC cells To further examine whether NLRP3-associated pyroptotic signaling contributes to the effects of tumor-cell-intrinsic ZNF831 on macrophage behavior, we treated ZNF831-overexpressing TNBC cells with the selective NLRP3 inhibitor MCC950 and then collected conditioned medium (CM) for downstream macrophage assays. Transwell migration assays showed that CM from ZNF831-overexpressing MDA-MB-231 and 4T1-luc cells enhanced RAW264.7 recruitment, whereas this effect was markedly attenuated when tumor cells were treated with MCC950 (Fig. 5A, B). Flow cytometric analysis showed that CM from ZNF831-overexpressing tumor cells increased the proportion of F4/80⁺CD86⁺ macrophages, whereas this increase was partially reversed by MCC950 treatment (Fig. 5C, D). Consistently, RT-qPCR and western blot analyses demonstrated that MCC950 attenuated the upregulation of M1-associated markers, including CD86, CD80, iNOS, and IL-6, as well as the protein expression of CD86 and iNOS (Fig. 5E, F). In parallel, CM from ZNF831-overexpressing tumor cells reduced the proportion of F4/80⁺CD206⁺ macrophages, while MCC950 partially restored this population (Fig. 5G, H). RT-qPCR and western blot analyses further showed that MCC950 partially restored M2-associated markers, including CD206, ARG1, CD163, and IL-10, together with the protein expression of CD206 and ARG1 (Fig. 5I, J). To verify that these macrophage-related effects were attenuated in parallel with suppression of pyroptotic signaling in tumor cells, we further examined key pyroptosis-associated readouts in the corresponding TNBC cells. LDH release assays showed that MCC950 attenuated the increase in lytic cell death associated with ZNF831 overexpression in both MDA-MB-231 and 4T1-luc cells (Fig. 5K). Western blot analysis showed that MCC950 reduced the ZNF831-induced increases in cleaved caspase-1 (p20) and GSDMD-NT (Fig. 5L). In addition, western blot analysis of concentrated culture supernatants and ELISA both demonstrated that MCC950 diminished the increase in extracellular IL-1β and IL-18 associated with ZNF831 overexpression (Fig. 5M, N). Together, these data indicate that pharmacologic inhibition of NLRP3 signaling attenuates the effects of ZNF831-overexpressing TNBC cells on macrophage recruitment and the associated shift toward a more pro-inflammatory, M1-like phenotype. Fig. 5MCC950 attenuates the macrophage-modulating effects of ZNF831-overexpressing TNBC cells. A, B Representative images and quantification of RAW264.7 transwell migration induced by conditioned medium from ZNF831-overexpressing MDA-MB-231 or 4T1-luc cells, with or without MCC950 treatment. C, D Representative flow cytometry plots and quantification of F4/80⁺CD86⁺ macrophages after stimulation with conditioned medium from the indicated tumor cell groups. E, F RT-qPCR and western blot analyses of M1-associated markers (CD86, CD80, iNOS, and IL-6) in RAW264.7 cells treated with conditioned medium from the indicated groups. G, H Representative flow cytometry plots and quantification of F4/80⁺CD206⁺ macrophages after stimulation with conditioned medium from the indicated tumor cell groups. I, J RT-qPCR and western blot analyses of M2-associated markers (CD206, ARG1, CD163, and IL-10) in RAW264.7 cells treated with conditioned medium from the indicated groups. K LDH release assays showing that MCC950 attenuated the increase in lytic cell death associated with ZNF831 overexpression in MDA-MB-231 and 4T1-luc cells. Asterisks indicate comparisons with the corresponding Vector groups, and number signs indicate comparisons with the corresponding ZNF831 groups at the same time point. L Western blot analysis of cleaved caspase-1 (p20) and GSDMD-NT in MDA-MB-231 and 4T1-luc cells under Vector, ZNF831, or ZNF831 + MCC950 conditions. M Western blot analysis of mature IL-1β and mature IL-18 in concentrated culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. N ELISA quantification of secreted IL-1β and IL-18 in culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. Data are presented as mean ± SD from three independent experiments. For panel K, statistical comparisons were performed at each time point by one-way ANOVA with Tukey’s multiple-comparisons test. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001 versus Vector; #p < 0.05, ##p < 0.01, ###p < 0.001 versus ZNF831 MCC950 attenuates the macrophage-modulating effects of ZNF831-overexpressing TNBC cells. A, B Representative images and quantification of RAW264.7 transwell migration induced by conditioned medium from ZNF831-overexpressing MDA-MB-231 or 4T1-luc cells, with or without MCC950 treatment. C, D Representative flow cytometry plots and quantification of F4/80⁺CD86⁺ macrophages after stimulation with conditioned medium from the indicated tumor cell groups. E, F RT-qPCR and western blot analyses of M1-associated markers (CD86, CD80, iNOS, and IL-6) in RAW264.7 cells treated with conditioned medium from the indicated groups. G, H Representative flow cytometry plots and quantification of F4/80⁺CD206⁺ macrophages after stimulation with conditioned medium from the indicated tumor cell groups. I, J RT-qPCR and western blot analyses of M2-associated markers (CD206, ARG1, CD163, and IL-10) in RAW264.7 cells treated with conditioned medium from the indicated groups. K LDH release assays showing that MCC950 attenuated the increase in lytic cell death associated with ZNF831 overexpression in MDA-MB-231 and 4T1-luc cells. Asterisks indicate comparisons with the corresponding Vector groups, and number signs indicate comparisons with the corresponding ZNF831 groups at the same time point. L Western blot analysis of cleaved caspase-1 (p20) and GSDMD-NT in MDA-MB-231 and 4T1-luc cells under Vector, ZNF831, or ZNF831 + MCC950 conditions. M Western blot analysis of mature IL-1β and mature IL-18 in concentrated culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. N ELISA quantification of secreted IL-1β and IL-18 in culture supernatants from MDA-MB-231 and 4T1-luc cells under the indicated conditions. Data are presented as mean ± SD from three independent experiments. For panel K, statistical comparisons were performed at each time point by one-way ANOVA with Tukey’s multiple-comparisons test. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001 versus Vector; #p < 0.05, ##p < 0.01, ###p < 0.001 versus ZNF831 ChIP, mRNA stability, and RIP analyses in MDA-MB-231 cells support multi-level regulation of NLRP3 by ZNF831 To further investigate how ZNF831 regulates NLRP3 expression, we examined whether ZNF831 associates with the NLRP3 promoter in the human TNBC cell line MDA-MB-231. Based on predicted C2H2 zinc finger binding sites within the human NLRP3 promoter, we designed ChIP-qPCR amplicons spanning three candidate regions (Fig. 6A). ChIP-qPCR performed in FLAG-ZNF831-overexpressing MDA-MB-231 cells showed enriched FLAG occupancy at Region 1, Region 2, and Region 3 compared with vector controls (Fig. 6B–D). In contrast, the RPL30 control locus did not show a comparable enrichment pattern (Fig. S4A). We next evaluated whether ZNF831 may also contribute to post-transcriptional regulation of NLRP3. In actinomycin D chase experiments, ZNF831 overexpression prolonged the apparent stability of NLRP3 mRNA in MDA-MB-231 cells (Fig. 6E). Consistent with this observation, RIP-qPCR demonstrated selective enrichment of NLRP3 mRNA in the FLAG-ZNF831 pull-down, whereas GAPDH mRNA did not show comparable enrichment (Figs. 6F and S4B). Corresponding HEK293T ChIP/RIP assays are provided as supplementary technical validation (Fig. S4C–H), but the primary mechanistic interpretation is based on the MDA-MB-231 data. Together, these results support a model in which ZNF831 promotes NLRP3 expression through combined promoter association and post-transcriptional interaction. Fig. 6ZNF831 associates with the NLRP3 promoter and mRNA in MDA-MB-231 cells. A Schematic of the human NLRP3 promoter showing predicted C2H2 zinc finger binding sites (Region 1–3) and the locations of ChIP-qPCR amplicons (GRCh38/hg38). B, D ChIP-qPCR analysis showing FLAG-ZNF831 enrichment at Region 1, Region 2, and Region 3 of the NLRP3 promoter in MDA-MB-231 cells. Data are expressed as % input. IgG served as a negative control and histone H3 served as a chromatin recovery control. E Actinomycin D chase assay showing prolonged apparent NLRP3 mRNA stability in ZNF831-overexpressing MDA-MB-231 cells. F RIP-qPCR analysis showing enrichment of NLRP3 mRNA in the FLAG-ZNF831 pull-down in MDA-MB-231 cells. Data are presented as mean ± SD from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 ZNF831 associates with the NLRP3 promoter and mRNA in MDA-MB-231 cells. A Schematic of the human NLRP3 promoter showing predicted C2H2 zinc finger binding sites (Region 1–3) and the locations of ChIP-qPCR amplicons (GRCh38/hg38). B, D ChIP-qPCR analysis showing FLAG-ZNF831 enrichment at Region 1, Region 2, and Region 3 of the NLRP3 promoter in MDA-MB-231 cells. Data are expressed as % input. IgG served as a negative control and histone H3 served as a chromatin recovery control. E Actinomycin D chase assay showing prolonged apparent NLRP3 mRNA stability in ZNF831-overexpressing MDA-MB-231 cells. F RIP-qPCR analysis showing enrichment of NLRP3 mRNA in the FLAG-ZNF831 pull-down in MDA-MB-231 cells. Data are presented as mean ± SD from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 ZNF831 restrains orthotopic TNBC growth and is accompanied by tissue-level inflammatory remodeling in vivo To assess the in vivo relevance of the proposed ZNF831–NLRP3-associated axis, we established an orthotopic syngeneic model by injecting 4T1-luc cells into the fourth mammary fat pad of BALB/c mice (Fig. 7A). ZNF831 overexpression significantly inhibited tumor growth, as reflected by tumor growth curves, representative bioluminescence imaging, quantitative analysis of the BLI results, excised tumor appearance, and final tumor weight (Fig. 7B–F). Intratumoral administration of MCC950 partially reversed these effects, supporting the involvement of NLRP3-related signaling in vivo. Histological analyses further supported remodeling of the tumor microenvironment. Immunohistochemistry showed increased CD86 staining and reduced CD206 staining in ZNF831-overexpressing tumors, whereas MCC950 treatment partially attenuated these changes (Fig. 7G, H). In parallel, double-immunofluorescence staining of tumor sections showed a higher abundance of F4/80+CD86+ cells and a lower abundance of F4/80+CD206+ cells in the ZNF831 group, both of which were partially reversed by MCC950 treatment (Fig. 7I–L). These tissue-level findings are consistent with a shift toward a more pro-inflammatory, M1-like macrophage phenotype in vivo, but do not constitute direct ex vivo quantification of tumor-associated macrophage subsets. Molecular analysis of resected tumors by RT-qPCR further showed that ZNF831 overexpression was associated with increased expression of pyroptosis-related genes and M1-like macrophage-associated markers, together with reduced expression of M2-like macrophage-associated markers, whereas MCC950 partially attenuated these changes (Fig. 7M–O). Taken together, these results support an association between ZNF831 overexpression, enhanced NLRP3-related pyroptotic signaling, and a more pro-inflammatory tumor microenvironment in vivo. Fig. 7ZNF831 suppresses TNBC growth in vivo and is associated with a more pro-inflammatory tumor microenvironment. A Schematic diagram of the orthotopic syngeneic model established by injection of 4T1-luc cells into the fourth mammary fat pad of BALB/c mice. B Tumor growth curves measured every 3 days. C Representative in vivo bioluminescence images of tumors in the indicated groups. D Quantitative analysis of relative tumor-associated BLI signal. ROI-derived CPS values were normalized to the mean value of the Vector group. E Representative images of excised tumors at the experimental endpoint. F Final tumor weights at the endpoint. G Representative H&E staining and immunohistochemistry for F4/80, CD86, and CD206 in tumor sections. H Quantification of CD86-positive cells and CD206-positive cells in tumor sections by immunohistochemistry. I Representative double-immunofluorescence staining of F4/80 and CD86 in tumor sections. J Representative double-immunofluorescence staining of F4/80 and CD206 in tumor sections. K Quantification of F4/80⁺CD86⁺ cells in tumor sections by immunofluorescence. L Quantification of F4/80⁺CD206⁺ cells in tumor sections by immunofluorescence. M RT-qPCR analysis of pyroptosis-related genes in resected tumors. N RT-qPCR analysis of M1-like macrophage-associated markers in resected tumors. O RT-qPCR analysis of M2-like macrophage-associated markers in resected tumors. Data are presented as mean ± SD (n = 5 mice per group). *p < 0.05, **p < 0.01, ***p < 0.001 ZNF831 suppresses TNBC growth in vivo and is associated with a more pro-inflammatory tumor microenvironment. A Schematic diagram of the orthotopic syngeneic model established by injection of 4T1-luc cells into the fourth mammary fat pad of BALB/c mice. B Tumor growth curves measured every 3 days. C Representative in vivo bioluminescence images of tumors in the indicated groups. D Quantitative analysis of relative tumor-associated BLI signal. ROI-derived CPS values were normalized to the mean value of the Vector group. E Representative images of excised tumors at the experimental endpoint. F Final tumor weights at the endpoint. G Representative H&E staining and immunohistochemistry for F4/80, CD86, and CD206 in tumor sections. H Quantification of CD86-positive cells and CD206-positive cells in tumor sections by immunohistochemistry. I Representative double-immunofluorescence staining of F4/80 and CD86 in tumor sections. J Representative double-immunofluorescence staining of F4/80 and CD206 in tumor sections. K Quantification of F4/80⁺CD86⁺ cells in tumor sections by immunofluorescence. L Quantification of F4/80⁺CD206⁺ cells in tumor sections by immunofluorescence. M RT-qPCR analysis of pyroptosis-related genes in resected tumors. N RT-qPCR analysis of M1-like macrophage-associated markers in resected tumors. O RT-qPCR analysis of M2-like macrophage-associated markers in resected tumors. Data are presented as mean ± SD (n = 5 mice per group). *p < 0.05, **p < 0.01, ***p < 0.001 Fig. 8Proposed working model of ZNF831-associated NLRP3 signaling, pyroptotic responses, and macrophage phenotypic remodeling in TNBC Proposed working model of ZNF831-associated NLRP3 signaling, pyroptotic responses, and macrophage phenotypic remodeling in TNBC ZNF831 is proposed to enhance NLRP3 expression through promoter association and interaction with NLRP3 mRNA, thereby promoting NLRP3-associated pyroptotic signaling, cytokine release, and a shift toward a more pro-inflammatory macrophage phenotype in the tumor microenvironment. MCC950 attenuated the tumor-suppressive and macrophage-related effects associated with ZNF831 overexpression in vivo; however, the present study did not determine whether these effects were mediated exclusively through tumor cells or also involved direct actions on tumor-infiltrating immune cells, including macrophages. Discussion In this study, we identified ZNF831 as a TNBC-associated regulator linked to a more inflammatory tumor microenvironment and favorable clinical outcome. Integrated bioinformatic analyses across scRNA-seq and bulk transcriptomic datasets showed that higher ZNF831 expression was associated with an M1-skewed macrophage signature, reduced M2-related features, and enrichment of pyroptosis-related inflammatory pathways. Functional studies in human and murine TNBC models further showed that ZNF831 overexpression was associated with increased NLRP3-related pyroptotic signaling, extracellular cytokine release, macrophage recruitment, and a shift toward a more pro-inflammatory, M1-like macrophage phenotype, whereas ZNF831 depletion produced the opposite trends. Together, these findings support a role for the ZNF831-NLRP3 axis in shaping a more inflammatory TNBC microenvironment (Fig. 8). Importantly, ZNF831 was not identified from a single correlational screen. Instead, it emerged from stepwise filtering across independent cohorts based on macrophage-related signatures, survival associations, and subtype specificity. This is relevant because TNBC is highly heterogeneous, and many immune-associated genes show context-dependent prognostic value across subtypes [17, 18]. In contrast, ZNF831 consistently linked favorable prognosis with a more inflammatory macrophage-associated profile in TNBC, suggesting that it may represent a biologically meaningful node connecting tumor-cell state to immune-context remodeling rather than a merely correlative biomarker. In this sense, our study places ZNF831 at the intersection of three processes that are particularly important for TNBC biology: inflammatory cell death, macrophage phenotypic remodeling, and immune-context remodeling. One important implication of our findings is that pyroptosis-related inflammatory signaling may help link tumor-cell responses to changes in macrophage phenotype within the TNBC microenvironment. Unlike non-inflammatory forms of tumor-cell death, pyroptotic signaling couples membrane rupture with release of inflammatory mediators [19, 20], thereby creating a local milieu that may influence neighboring immune cells [8, 21, 22]. In our system, ZNF831 overexpression was associated not only with increased cleavage of caspase-1 and GSDMD, but also with increased extracellular release of mature IL-1β and IL-18, together with enhanced macrophage recruitment and a shift toward M1-like features. Although our data do not prove that any single mediator is solely responsible for this macrophage response, they are consistent with a model in which pyroptosis-related inflammatory outputs contribute to a more immunostimulatory microenvironment. At the same time, the macrophage findings should be interpreted with appropriate caution. Macrophage states in vivo are increasingly understood as existing along a dynamic continuum rather than as two rigidly separated categories [10]. Accordingly, we intentionally adopted more cautious terminology such as “M1-like phenotype,” “M2-related features,” and “phenotypic remodeling,” rather than implying absolute polarization into fixed macrophage subsets. This is particularly important in TNBC, where macrophage composition and function may vary across disease stage, stromal context, and immune architecture. Therefore, our bioinformatic, in vitro, and in vivo data are best understood as supporting a shift toward a more pro-inflammatory macrophage state associated with ZNF831, rather than proving binary subtype conversion. Interestingly, our previous work showed that ZNF831 promotes apoptosis in breast cancer through repression of the STAT3/Bcl2 axis [15]. In contrast, the present study in TNBC supports a role for ZNF831 in NLRP3-associated pyroptotic signaling, as reflected by increased cleaved caspase-1 (p20), GSDMD-NT, LDH release, and extracellular release of mature IL-1β and IL-18. These observations may reflect context-dependent engagement of distinct programmed cell-death pathways rather than a direct contradiction. Increasing evidence indicates substantial crosstalk between apoptosis and pyroptosis, and the dominant death phenotype may vary depending on cell type, inflammatory context, upstream stimuli, and gasdermin expression status [23]. In other words, ZNF831 may not function as a dedicated inducer of only one death program, but rather as a regulator whose downstream consequences depend on the molecular landscape of the tumor cells under study. This interpretation is also consistent with the fact that the present work focused specifically on TNBC models under conditions in which inflammasome-related signaling was strongly engaged. However, because we did not systematically assess apoptotic morphology or apoptosis-specific markers in the current study, we cannot determine whether apoptosis coexists with pyroptosis in ZNF831-enriched TNBC cells, nor can we define a mechanistic transition between these two death programs. Clarifying whether ZNF831 coordinates multiple programmed cell-death pathways in a context-dependent manner will require dedicated future investigation. Our data also suggest a potentially interesting regulatory mode by which ZNF831 may sustain NLRP3 expression. The newly added ChIP-qPCR, actinomycin D chase, and RIP-qPCR data in MDA-MB-231 cells strengthen the physiological relevance of the ZNF831–NLRP3 interaction by showing promoter association, prolonged apparent NLRP3 mRNA stability, and selective enrichment of NLRP3 mRNA in a human TNBC context. This is noteworthy because KRAB-domain zinc finger proteins are classically regarded as transcriptional repressors through KAP1/TRIM28-associated mechanisms [24–26], yet increasing evidence indicates that members of this family can function in a context-dependent manner [27, 28]. In this setting, ZNF831 may participate in NLRP3 regulation through a combination of promoter association and post-transcriptional interaction, thereby helping sustain inflammasome-related signaling in TNBC. At the same time, our mechanistic data should not be interpreted as fully resolving the precise mode of NLRP3 regulation. Although the MDA-MB-231-based ChIP-qPCR and RIP-qPCR experiments improve biological relevance compared with a heterologous overexpression system alone, promoter-reporter assays and nascent transcript or pre-mRNA analyses will still be needed to rigorously define whether ZNF831 directly activates NLRP3 transcription and to further separate transcriptional from post-transcriptional effects. Likewise, additional mechanistic work will be needed to determine whether ZNF831 engages co-activator complexes, antagonizes constitutive repressors, or associates with NLRP3 mRNA through direct or indirect interactions. Thus, our current data support a model involving promoter association and post-transcriptional interaction, but do not yet define its full biochemical basis. Several limitations should also be acknowledged. First, our current evidence for the causal link between tumor-cell-intrinsic ZNF831/NLRP3 signaling and macrophage recruitment or phenotypic remodeling is based primarily on pharmacologic inhibition with MCC950. Genetic loss-of-function studies targeting NLRP3, CASP1, or GSDMD in tumor cells, as well as cytokine-neutralization experiments for IL-1β and IL-18, will be needed to more rigorously define the contribution of each node in this signaling cascade. Second, because MCC950 was administered intratumorally in vivo, we did not directly dissect its cell type-specific effects within the tumor microenvironment. In particular, we did not isolate tumor-associated macrophages or perform ex vivo analyses to determine whether MCC950 directly affected infiltrating macrophages in vivo. Therefore, although our data support the involvement of NLRP3-related signaling in the ZNF831-associated antitumor phenotype, the relative contributions of tumor-cell-intrinsic versus macrophage-intrinsic NLRP3 signaling remain to be further defined. Third, the in vitro macrophage experiments were performed in RAW264.7 cells rather than primary bone marrow-derived macrophages. Although RAW264.7 cells continue to be used in recent breast cancer and triple-negative breast cancer studies as a tractable in vitro model for mechanistic investigation of tumor–macrophage interactions [29, 30], they do not fully recapitulate the heterogeneity and functional plasticity of primary macrophages. In the present study, the macrophage-related conclusions are supported not only by RAW264.7-based in vitro assays but also by in vivo findings in an immunocompetent orthotopic TNBC model. Nevertheless, validation in primary macrophages, including bone marrow-derived macrophages, will be important in future work to further strengthen physiological relevance. Finally, while our work focused primarily on macrophages as the dominant immune population linked to ZNF831 in TNBC, additional studies will be needed to determine how this axis influences other immune compartments, including adaptive immunity [21, 31, 32]. Broader validation across TNBC models and clinical cohorts will also be important [33, 34]. From a translational perspective, the ZNF831–NLRP3 axis may be of interest as a biomarker-associated pathway for further investigation in TNBC. Because ZNF831 is downregulated in breast cancer, future studies may explore whether restoring ZNF831 expression or modulating downstream pyroptosis-related signaling could help reshape poorly inflamed TNBC microenvironments [17, 18, 35, 36]. However, the feasibility, safety, and therapeutic tractability of such strategies remain to be established [37, 38]. At the current stage, the main translational value of our findings lies less in proposing an immediate therapeutic strategy and more in providing a rationale for stratification and mechanism-guided follow-up studies in immunologically heterogeneous TNBC. In conclusion, our findings identify ZNF831 as a TNBC-associated regulator linked to enhanced NLRP3-related pyroptotic signaling, a more pro-inflammatory macrophage phenotype, and tumor suppression. By integrating bioinformatic analyses with in vitro and in vivo experiments, this study supports a role for the ZNF831–NLRP3 axis in shaping the TNBC microenvironment. These results provide a rationale for further investigation of this pathway as a candidate biomarker-associated pathway and for mechanism-guided follow-up studies in TNBC. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Tables S1–S3 and Supplementary Figures S1–S4. (DOCX). Supplementary Material 1. Supplementary Tables S1–S3 and Supplementary Figures S1–S4. (DOCX). Supplementary Material 2. Full uncropped Western blot images for Figs. 2, 4 and 5. (PDF). Supplementary Material 2. Full uncropped Western blot images for Figs. 2, 4 and 5. (PDF).