Article Text

Original research
Tumor cell-released autophagosomes (TRAPs) induce PD-L1-decorated NETs that suppress T-cell function to promote breast cancer pulmonary metastasis
  1. Xiaohe Zhou1,
  2. Chengdong Wu1,
  3. Xuru Wang1,
  4. Ning Pan1,
  5. Xiaotong Sun2,
  6. Bohao Chen1,
  7. Shiya Zheng3,
  8. Yiting Wei1,
  9. Jing Chen1,
  10. Yuyang Wu1,
  11. Fengjiao Zhu1,
  12. Jinpeng Chen4,
  13. Huabiao Chen5 and
  14. Li-xin Wang1
  1. 1Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Microbiology and Immunology, School of Medicine, Southeast University, Nanjing, Jiangsu, China
  2. 2Department of Laboratory Medicine, Yantai Yuhuangding Hospital, Yantai, Shandong, China
  3. 3Department of Oncology, Southeast University Zhongda Hospital, Nanjing, Jiangsu, China
  4. 4Department of general surgery, Southeast University Zhongda Hospital, Nanjing, Jiangsu, China
  5. 5Institute of Biomedical Engineering and Technology, School of Medicine, Ningbo University, Ningbo, China
  1. Correspondence to Dr Li-xin Wang; lxwang{at}
  • XZ, CW, XW and NP are joint first authors.


Background Lung metastasis is the primary cause of breast cancer-related mortality. Neutrophil extracellular traps (NETs) are involved in the progression of breast cancer. However, the mechanism of NET formation is not fully understood. This study posits that tumor cell-released autophagosomes (TRAPs) play a crucial role in this process.

Methods TRAPs were isolated from breast cancer cell lines to analyze their impact on NET formation in both human and mouse neutrophils. The study used both in vitro and in vivo models, including Toll-like receptor 4 (TLR4/) mice and engineered breast cancer cell lines. Immunofluorescence, ELISA, Western blotting, RNA sequencing, and flow cytometry were employed to dissect the signaling pathways leading to NET production and to explore their immunosuppressive effects, particularly focusing on the impact of NETs on T-cell function. The therapeutic potential of targeting TRAP-induced NETs and their immunosuppressive functions was evaluated using DNase I and αPD-L1 antibodies. Clinical relevance was assessed by correlating circulating levels of TRAPs and NETs with lung metastasis in patients with breast cancer.

Results This study showed that TRAPs induced the formation of NETs in both human and mouse neutrophils by using the high mobility group box 1 and activating the TLR4-Myd88-ERK/p38 signaling axis. More importantly, PD-L1 carried by TRAP-induced NETs inhibited T-cell function in vitro and in vivo, thereby contributing to the formation of lung premetastatic niche (PMN) immunosuppression. In contrast, Becn1 KD-4T1 breast tumors with decreased circulating TRAPs in vivo reduced the formation of NETs, which in turn attenuated the immunosuppressive effects in PMN and resulted in a reduction of breast cancer pulmonary metastasis in murine models. Moreover, treatment with αPD-L1 in combination with DNase I that degraded NETs restored T-cell function and significantly reduced tumor metastasis. TRAP levels in the peripheral blood positively correlated with NET levels and lung metastasis in patients with breast cancer.

Conclusions Our results demonstrate a novel role of TRAPs in the formation of PD-L1-decorated NETs, which may provide a new strategy for early detection and treatment of pulmonary metastasis in patients with breast cancer.

  • Breast Cancer
  • Immune modulatory
  • Neutrophil

Data availability statement

No data are available. Not applicable.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Although it is recognized that neutrophil extracellular traps (NETs) play significant roles in the progression of breast cancer and lung metastasis, the detailed mechanisms of NET formation and their specific role in immunosuppression at metastatic sites have not been fully elucidated.


  • Our research for the first time reveals that tumor cell-released autophagosomes (TRAPs) trigger NET formation through the high mobility group box 1 and Toll-like receptor 4 -MyD88-ERK/p38 pathway. Moreover, we demonstrate that PD-L1 carried by TRAP-induced NETs plays a vital role in suppressing T-cell function, thereby fostering an immunosuppressive lung premetastatic niche.


  • The study highlights the role of TRAP-induced PD-L1-decorated NETs in breast cancer progression, serving as potential biomarkers for predicting lung metastasis and offering a novel strategy for the early detection and treatment of lung metastasis in patients with breast cancer.


Pulmonary metastasis stands as the most prevalent form of distant metastasis and a principal cause of mortality among patients with breast cancer (BC).1 Lung metastasis of BC is a multistage process regulated by multiple factors. It has been reported that primary tumor can promote the formation of a premetastatic niche (PMN) by releasing tumor-derived soluble factors, which is characterized by immunosuppression, inflammation, and angiogenesis at the metastasis site to provide fertile soil for further metastasis of tumor cells.2–6

Neutrophils are the body’s initial defense against pathogen invasion but can serve as a protumoral component of the immune system in the tumor microenvironment (TME) as they are closely associated with tumor occurrence and development.7 Recent research has demonstrated that neutrophils release neutrophil extracellular traps (NETs) that play significant roles in the formation of the PMN and the coordination of the entire tumor metastatic process.8 9 NETs are mesh-like structures composed of DNA, histones, and cytotoxic granule-derived proteins10 and can capture and kill pathogens during infection.11 12 Recent studies have documented the presence of NETs in various tumor distal organs, including the liver, lungs, and omentum.13–15 NETs in the bloodstream have been shown to be capable of capturing dispersed cancer cells and facilitating their metastasis and dissemination.16 Additionally, NETs can promote cancer metastasis by binding to the receptor CCDC25 on the surface of cancer cells.15 However, the mechanisms of NET formation remain to be investigated.

Previous studies have shown that tumor cells produce certain mediators, such as secreted factors and extracellular vesicles (EVs), and interact with various cells in potential metastatic organs to promote PMN formation.17 In our previous investigations, we identified tumor-released autophagosomes (TRAPs) in tumor cell supernatants as well as in the malignant effusions and ascites of patients with cancer.18 19 TRAPs modulate the immunosuppressive activity of B cells,19 macrophages,20 and CD4+ T cells21 within the TME. Furthermore, TRAPs can promote neutrophil apoptosis which inhibits T-cell proliferation in a cell-contact and ROS-dependent manner.22 Additionally, TRAPs can influence lung PMN formation via the activation of pulmonary fibroblasts in BC.23

In this study, we demonstrate that TRAPs induced the formation of NETs through the activation of the neutrophil Toll-like receptor 4 (TLR4-Myd88-ERK/p38) signaling pathway mediated by high mobility group box 1 (HMGB1). PD-L1 carried by TRAP-induced NETs suppressed T-cell functionality, thereby creating an immunosuppressive premetastatic environment, which further promoted BC pulmonary metastasis. Furthermore, αPD-L1 combined with DNase I for NET degradation significantly restored T-cell function. Our study reveals a novel role for TRAPs as inducers of the release of PD-L1-decorated NETs that suppress T-cell function, suggesting that TRAPs and NETs may serve as potential biomarkers for the early detection and treatment of lung metastasis in patients with BC.

Materials and methods


Between January 2023 and January 2024, peripheral blood samples were collected from 105 patients with BC at the Affiliated Yantai Yuhuangding Hospital of Qingdao University. The inclusion criteria were as follows: pathologically confirmed diagnosis of BC and signed informed consent. The exclusion criteria were as follows: age under 18 or over 80 years; pregnant women; concurrent or previous history of other solid malignancies or hematological malignancies besides BC; coexisting autoimmune diseases; or long-term use of immunosuppressive drugs. Concurrently, peripheral blood samples from 42 healthy donors of varying ages were also collected for control purposes.


Specific Pathogen-Free (SPF) grade female C57BL/6 and BALB/c mice aged 6–8 weeks were sourced from Yangzhou University’s Center for Comparative Medicine. Additionally, Tlr4-deficient (Tlr4−/−) mice were obtained from the Nanjing Biomedical Research Institute. All mice were housed under SPF conditions under a standard light/dark cycle. All animal experiments were approved by the Southeast University Animal Care and Use Committee (No. 20210910045) and conducted in accordance with ethical regulations.

Cell lines and culture conditions

The human MDA-MB-231 BC cell line and the mouse 4T1 BC cell line were procured from the Shanghai Institute of Cell and Biochemistry of the Chinese Academy of Sciences. MDA-MB-231 cells were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS), penicillin (100 U/mL, Gibco), and streptomycin (0.1 mg/mL, C0222, Beyotime) in a 37°C and 5% CO2 incubator. 4T1 cells were maintained in RPMI-1640 medium containing 10% FBS. In our laboratory, we engineered Becn1 knockdown in 4T1 cells (Becn1 KD) using a lentivirus encoding a shRNA that targets the Becn1 gene sequence (5′-GCGGGAGUAUAGUGAGUUUTT-3’). Corresponding non-targeted control cells (Becn1 NC) were created using a non-specific vector with a scrambled sequence (5′-TTCTCCGAACGTGTCACGTAA-3′) provided by Hanbio Biotechnology, Shanghai, China. Similarly, Hmgb1 knockdown cells (Hmgb1-KD) were developed using a lentivirus encoding an shRNA targeting the Hmgb1 gene (5′-GATATGGCAAAGGCTGACA-3′) while corresponding non-targeted control cells (Hmgb1-NC) were prepared using a scrambled shRNA sequence (5′-TTCTCCGAACGTGTCACGT-3′) sourced from GenePharma (Shanghai, China). EGFP-LC3 fusion-expressing 4T1 cells were constructed by transfecting with EGFP-LC3 and luciferase-expressing lentiviruses, which were purchased from Hanbio Biotechnology (Shanghai, China). PD-L1 knockdown in 4T1 cells was achieved using a lentivirus expressing PD-L1-targeting shRNA (sequence 5′-TGACGTTGCTGCCAT-3′), sourced from GENECHEM CO. (Shanghai, China). Mycoplasma contamination was routinely checked every two weeks using the EZ-PCR Mycoplasma Detection Kit.


TLR2 inhibitor (TLR2-IN-C29), TLR4 inhibitor (TLR4-IN-C34), ERK inhibitor (SCH772984), and P38 inhibitor (BIRB 796) compounds were purchased from Selleck (Shanghai, China). The used MyD88 inhibitor (ST 2825) was sourced from MCE (Shanghai, China).

Neutrophil isolation

Neutrophils were isolated from human peripheral blood using Histopaque 1.119 (11191, Sigma) and 1.077 (10771, Sigma) density gradient centrifugation, and from mouse bone marrow using the Neutrophil Isolation Kit (130-097-658, Miltenyi Biotec). Purity (>97%) was confirmed by flow cytometry with anti-CD66b (392904, BioLegend) for human neutrophils, and anti-CD11b (101206, BioLegend) and anti-Ly6G (127608, BioLegend) for mouse neutrophils. Procedures were conducted in accordance with respective reagent manuals.

TRAP purification and characterization

When cells were healthy and 70%–90% confluent as noted previously,24 TRAPs were harvested from MDA-MB-231 and 4T1 cells. Briefly, culture supernatants were collected and centrifuged at 3000 rpm to remove dead cells and fragments. The supernatant fraction following centrifugation at 12 000 rpm was then used to harvest TRAPs secreted by BC cells. TRAPs were harvested from the plasma of patients with BC using a similar method. Then, TRAPs were stained with an Alexa Fluor 647 anti-LC3 antibody (ab225383, Abcam), and the purity of TRAPs was analyzed by flow cytometry. TRAPs morphology was detected by transmission electron microscopy (TEM). The LC3 protein in TRAPs was detected by Western blotting.

In vitro neutrophil stimulation

Neutrophils were seeded in 24-well plates (5×105 cells/well) and incubated for 1 hour. Subsequently, the cells were treated with 100 nM phorbol 12-myristate 13-acetate (PMA) (P1585, Sigma-Aldrich) and 10 µg/mL TRAP for 3 hours. Concurrently, some wells were treated with either 10 µM GSK484 (HY-100514, MCE, USA) or 10 U DNase I (EN0521, Thermo Scientific) while cells in the control group received an equivalent volume of PBS. After the treatment period, the supernatant was collected to analyze free DNA content. The presence of characteristic NETs molecules, such as neutrophil elastase (NE), myeloperoxidase (MPO), and citrullinated histone H3 (H3Cit), was determined through immunofluorescence (IF) staining.

Quantification of free dsDNA

We employed the Quant-iT PicoGreen dsDNA Assay Kits and dsDNA Reagents (P11495, Invitrogen) to quantify free DNA in cell supernatants. Following the kit instructions, diluted samples and standards, in 50 µL volumes, were added to a 96-well plate, each followed by 50 µL of PicoGreen reagent. Standards were prepared in a concentration range from 0 to 1000 ng/mL. Following a brief incubation in a dark environment, fluorescence was measured at an excitation of 480 nm and an emission of 520 nm.

Detection of MPO-DNA complexes

To quantify NETs, a modified capture ELISA method was used for detecting MPO-DNA complexes, as outlined in previous studies.25 26 The procedure involved coating a 96-well plate with 5 µg/mL anti-MPO monoclonal antibody (SAB1409321, Sigma) overnight at 4°C. After blocking with 1% BSA, test samples and a peroxidase-labeled anti-DNA monoclonal antibody (11774425001, Roche) were added to the wells and incubated for 2 hours at room temperature. Following three times washes with PBS, the ABTS peroxidase substrate from the kit was added, and the plate was incubated for 40 min at 37°C in darkness. Absorbance readings were then taken at 405 nm.

Purification of NETs

Neutrophils, isolated using the previously described method, were stimulated with TRAPs in Phenol-red-free RPMI-1640 containing 1% FBS for 6 hours. After stimulation, the supernatant was removed. Then, plate-bound NETs and neutrophils were washed with cold PBS, centrifuged at 450×g to separate the neutrophils, yielding a supernatant enriched for NETs. This NETs-rich supernatant was further centrifuged at 14,000×g and resuspended in cold PBS to a final concentration of 2×107 neutrophils per 100 µL. NETs morphology was assessed using SYTOX Green (S7020, Thermo Fisher Science), and the prepared NETs were used for subsequent experiments.

Scanning electron microscopy

Neutrophils were placed onto cover slides and treated with 100 nM PMA, TRAPs, or were left untreated for 3 hours. The samples were then processed for electron microscopy as previously described. This involved fixing the samples overnight in 2.5% glutaraldehyde, followed by washing with PBS and incubation in 1% cesium tetroxide. The samples were dehydrated using a graded series of ethanol, subjected to critical point drying, and coated with 2 nm of platinum. Finally, a 5 nm carbon layer was applied before examination under a JEOL-7900F scanning electron microscope.

IF staining

For cellular IF analyses, aseptic glass slides pretreated with polylysine were placed in a 24-well plate. Cells were then added at a density of 5×105 per well and stimulated with PMA or TRAP as previously described. After a 3-hour incubation, the supernatant was removed, and cells were fixed with 0.5 mL of 4% paraformaldehyde. The fixed cells were washed three times with PBS, permeabilized with 0.5% TritonX-100 in PBS, and blocked with 5% goat serum at 37°C for 30 min. Primary antibodies including anti-H3Cit (1:400, ab5103, Abcam, Cambridge, UK), anti-MPO (1:400, SAB1409321, Sigma), and anti-CD274 (PD-L1, B7-H1) (1:500, 14-5982-82, Thermo Fisher Scientific) were applied, followed by incubation at 37°C for 1 hour. Postwashing, secondary antibodies including Alexa Fluor 647 Donkey Anti-Rabbit IgG H&L (1:1000, ab150075, Abcam), Alexa Fluor 488 Goat Anti-Mouse IgG H&L (1:1000, RS3208, ImmunoWay), and Dylight 594 Goat Anti-Rat lgG (1:1000, RS23440, ImmunoWay) were added and samples were incubated for 1 hour. Finally, after a further PBS wash, DAPI (1:2000) was used to stain cells for 5 min while protected from light, and after washing, the slides were imaged using an FV3000 laser confocal microscope.

For tissue IF analyses, lung samples were fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin. Thin slices (4–5 µm) were prepared and mounted on glass slides. The slides were deparaffinized with xylene, rehydrated using a graded alcohol series, and antigen retrieval was performed in citrate buffer. After blocking with 5% goat serum, slides were incubated with primary antibodies (anti-H3Cit, anti-MPO, and anti-CD274). This was followed by incubation with secondary antibodies (conjugated to AlexaFluor 647, AlexaFluor 488, and Dylight594). After incubation, nuclei were stained with DAPI. The slides were then examined under an FV3000 laser confocal microscope to analyze NETs.

T cell proliferation and activation assays

For T cell activation assays, mouse splenocytes were seeded in 48-well plates coated overnight with 2 mg/µL anti-CD3 and anti-CD28 antibodies (eBioscience). Depending on whether NETs were present, different treatments were applied: either 20 µg/mL of αPD-L1 neutralizing monoclonal antibody (BioLegend), DNase I alone, or a combination of anti-PD-L1 antibody and DNase I. During the final 5 hours of coculture, a protein transport inhibitor (eBioscience) was added. The secretion of IFN-γ and the proliferation of T cells were then assessed using flow cytometry.

To detect T-cell function in murine lung tissue, lung tissues were harvested following euthanasia, minced, and incubated for 1 hour at 37°C with DMEM containing collagenase IV (1 mg/mL), hyaluronidase (5 ku/mL), and DNase I (20 U/mL) (all from Sigma). The mixture was then passed through a 70 µm filter and the resulting cell suspension was centrifuged followed by layering over mouse lymphocyte separation solution (Dakewe) and centrifugation at 600×g for 15 min at room temperature. T cells isolated from lung tissues were stimulated using the protein transport inhibitor (00-4980-93, eBioscience, California, USA) for 5 hours before collection. Flow cytometry was used to quantify IFN-γ+ and Ki67+ T cells.

Flow cytometry

A single-cell suspension was prepared and stained with Fixable Viability Stain 520 or Fixable Viability Stain 780 (eBioscience) to exclude dead cells. Cells were then blocked with Human or Mouse FcR Blocker at 4°C for 15 min. For human samples, used antibodies included PE-CD66b (392904, BioLegend) and Alexa Fluor 647-LC3B (ab225383, Abcam). For mouse samples, used antibodies included CD11b (101206, BioLegend), Ly6G (127614, BioLegend), HSP27 (49283-1, Signalway Antibody), HSP60 (ab190828, Abcam), HSP70 (ab181606, Abcam), HSP90 (ab79849, Abcam), HMGB1 (RG65863, Arigo biolaboratories), PE-IFN-γ (505808, BioLegend), PE-Ki-67 (652404, BioLegend), Alexa Fluor 700-CD45 (56-0451-82, eBioscience), Percp-Cy5.5-CD3 (100326, BioLegend), APC-CD4 (116014, BioLegend), FITC-CD4 (100405, BioLegend), APC-CD8a (100712, BioLegend), PE-PD-L1 (124308, BioLegend), Alexa Fluor 647 Donkey Anti-Rabbit IgG H&L (ab150075, Abcam), and Alexa Fluor 488 Goat Anti-Mouse IgG H&L (RS3208, ImmunoWay). A Fixation/Permeabilization Kit (554714, BD) was used for cytoplasmic protein staining, and a Transcription Factor Buffer Set (562574, BD) was used for nuclear staining. Flow cytometry was performed using FACS Calibur, Thermo AttuneNxt, or Cytek Northernlight systems, and data were analyzed using FlowJo V.10 software (TreeStar, Ashland, USA).

Western blotting

Total protein extraction was performed using RIPA buffer (WB3100, New Cell & Molecular, China) with protease and phosphatase inhibitors (P002, New Cell & Molecular). Proteins were then separated via 10% or 15% PAGE and transferred onto PVDF membranes (Millipore) with the rapid transfer buffer from New Cell & Molecular Biotech. For blocking, blots were incubated with 5% BSA or skim milk powder. Used primary antibodies included anti-LC3 (Sigma-Aldrich), anti-LC3B (Cell Signaling Technology), anti-Beclin1 (11306-1-AP, Proteintech), anti-β-tubulin (10094-1-AP, Proteintech), and anti-GAPDH (60004-1-Ig, Proteintech), as well as antibodies specific for P38, phosphor-P38, ERK, and phosphor-ERK (all from Cell Signaling Technologies). Secondary antibodies included goat anti-mouse and anti-rabbit IgG HRP, both from Proteintech.

Animal experiments

To develop a mouse model with reduced TRAP release, Becn1 KD-4T1 or Becn1 NC-4T1 cells (5×105 cells/100 µL) were injected into the right fourth mammary gland fat pad. Plasma samples were collected on days 7, 10, and 14 for TRAP quantification via flow cytometry. NETs levels in plasma and lung tissue were measured on day 10 using flow cytometry and tissue IF, respectively. In addition, lung T-cell functionality, including IFN-γ secretion and proliferation, was assessed. On day 35, lung tissues were fixed with Bouin’s solution for the counting of pulmonary nodules and the observation of spontaneous metastasis.

In the mouse intravenous TRAP model, 10 µg of TRAPs was administered intravenously every other day for a total of five administrations. NET levels in plasma and lung tissue were measured on day 10. Lung T-cell function was evaluated via flow cytometry. To assess pulmonary metastasis, Luc-4T1 cells (1×105 cells/100 µL) were injected into the caudal vein. After 25 days, pulmonary metastatic foci were imaged in vivo, and lung nodules were examined after mice had been euthanized.

In the therapeutic model, mice received intravenous injections of TRAPs and in situ 4T1 cells (5×105 cells/100 µL). Each mouse was intraperitoneally injected with 200 µg of an αPD-L1 antibody (BE0101, BioXCell) or control IgG2a (BE0089, BioXCell) on days 5, 7, and 9 for a total of three injections. For DNase I treatment, 75 U of DNase I was intraperitoneally administered each day. NET levels in the plasma and pulmonary T-cell function were assessed on day 10. Pulmonary metastasis was evaluated by injecting 4T1 cells into the tail vein and imaging metastatic lesions in vivo after 25 days. After euthanasia, lung tissue was fixed with Bouin’s solution to count pulmonary nodules.

Statistical analysis

Statistical analyses and figure preparation were performed using GraphPad Prism V.9.0. Experimental data and statistical significance of differences were analyzed using unpaired Student’s t-test or one-way analysis of variance. Tumor growth curves were analyzed using the unpaired Mann-Whitney test. Binary logistic regression analyses were employed to identify independent risk factors. The accuracy of variables in predicting lung metastasis was evaluated using receiver-operating characteristic (ROC) curves and by calculating the area under the curve (AUC). Data are presented as means±SEM from at least three independent experiments. A p value of less than 0.05 was considered statistically significant (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, NS: not significant).


TRAPs induce NET formation and enhance the pulmonary metastasis in BC

Our study aimed to elucidate whether TRAPs could instigate NET generation within these niches. To that end, we engineered and characterized a 4T1 BC cell line in which the Becn1 gene had been knocked down (Becn1 KD-4T1), alongside a control cell line (Becn1 NC-4T1) incorporating a non-targeting shRNA sequence. An in situ BC tumor-bearing model was then established (figure 1A). Becn1 KD-4T cells curtailed intracellular LC3-II accumulation and markedly decreased TRAP release (online supplemental figure S1A). Despite slower tumor growth in Becn1 KD-4T1 mice, tumor sizes matched those in Becn1 NC-4T1 mice within 14 days (online supplemental figure S1B). Intriguingly, TRAP levels in the plasma of Becn1 NC-4T1 mice were substantially higher than in Becn1 KD-4T1 mice (online supplemental figure S1C).

Supplemental material

Figure 1

Tumor cell-released autophagosomes (TRAPs) induce neutrophil extracellular trap (NET) formation and enhance the pulmonary metastasis of breast cancer. (A) Establishment of a tumor-bearing mouse model with reduced TRAP release. Experimental groups included the TF, Becn1 NC (negative control), Becn1 KD (knockdown), and Becn1 KD+TRAPs groups. (B) Plasma levels of dsDNA and myeloperoxidase-DNA (MPO-DNA) complexes were measured in the four groups on day 10. (C,D) Immunofluorescence staining for DNA (blue), MPO (green), and H3Cit (red) in lung tissues from each group. (E) Representative images of pulmonary surface nodules on day 35 (n=5 mice/group). (F) Schematic overview of the TRAP tail vein injection mouse model (n=5). Mice received intravenous 4T1-TRAP injections every 2 days for a total of five injections. (G,H) Immunofluorescence staining for DNA (blue), MPO (green), and H3Cit (red) in lung tissues from the TRAP and control groups on day 10. (I) Plasma levels of dsDNA and MPO-DNA complexes were measured on day 10. (J) Western blot analysis of H3Cit protein levels in lung tissues from the TRAP and control groups on day 10. (K,L) Lung colonization was analyzed following the intravenous injection of Luc-4T1 cells. Representative bioluminescence imaging (BLI) (K) and images of pulmonary surface nodules (L) are shown for day 35 (n=5 mice/group). Data (mean±SEM) represent three independent experiments (*p< 0.05; **p< 0.01; ***p< 0.001; ****p< 0.0001; NS, not significant).

We initially isolated TRAPs from 4T1 cell culture supernatants, characterized by the presence of the mature autophagy marker LC-3II and a double-membrane structure measuring 300–900 nm in diameter (online supplemental figure S1D–F). DiR-labeled TRAPs were intravenously injected into mice and detected in lung tissue 24 hours afterwards (online supplemental figure S1G), signifying their ability to reach lung tissues. Moreover, on day 10 after the subcutaneous injection of EGFP-LC3-4T1 cells into the right fourth mammary gland fat pad, green fluorescence can be observed in the lung tissue of tumor-bearing mice (online supplemental figure S1H,I). Based on prior research, tumor-bearing mice were deemed to be in a premetastatic phase within 14 days.23 Consequently, we assessed NET levels in tumor-bearing mice on day 10, examining the involvement of NETs in PMN formation. The results showed significantly lower plasma dsDNA and MPO-DNA complex levels in Becn1 KD-4T1 mice, which surged following intravenous TRAP injection (figure 1B). Lung tissue sections stained for the NET markers H3Cit and MPO revealed pronounced NET formation in Becn1 NC-4T1 tumor-bearing mice compared with Becn1 KD-4T1 mice. However, NET formation was particularly heightened in Becn1 KD-4T1 mice after intravenous TRAP injection (figure 1C,D).

Flow cytometry analyses demonstrated a lower percentage of CD45+ cell-infiltrating neutrophils in the lung tissues of Becn1 KD-4T1 mice compared with Becn1 NC-4T1 mice. This percentage rose following intravenous TRAP supplementation (online supplemental figure S1J,K). Likewise, spontaneous lung metastasis occurrence was notably reduced in Becn1 KD-4T1 mice but significantly increased following TRAP supplementation (figure 1E). Intravenous injection of TRAPs into mice led to elevated NET levels in both plasma and lung tissue after five alternate-day injections (figure 1F–H), with higher plasma dsDNA and MPO-DNA complex levels (figure 1I) and increased neutrophil infiltration in the lung tissue (online supplemental figure S1I). Western blotting analyses of the lung tissue revealed significantly higher levels of H3Cit, a characteristic component of NETs, after TRAP supplementation (figure 1J). Increased pulmonary metastasis was also observed after TRAP supplementation in luc-4T1 tumor-bearing mice (figure 1K,L). Collectively, these findings suggest that TRAPs can induce NET formation in vivo, contribute to premetastatic lung niche development and enhance BC lung metastasis.

TRAPs induce NET production by neutrophils in vitro

Neutrophils were isolated from healthy human peripheral blood and mouse bone marrow. High activity and purity of isolated human and mouse neutrophils were achieved for subsequent experimental studies (figure 2A, B, F,G).

Figure 2

Tumor cell-released autophagosomes (TRAPs) can induce human and mouse neutrophils to form neutrophil extracellular traps (NETs) in vitro. (A,F) The purity of neutrophils obtained from human peripheral blood and mouse bone marrow was analyzed by FCM. (B,G) The purity of neutrophils obtained from human peripheral blood and mouse bone marrow by Wright-Giemsa staining. (C,H) Human and mouse neutrophils were treated with phorbol 12-myristate 13-acetate (PMA) (100 nM), TRAPs derived from the human MDA-MB-231 breast cancer cell line (10 µg/mL), or TRAPs from the mouse 4T1 breast cancer cell line (10 µg/mL) for 3 hours. The morphology of NETs was observed using a scanning electron microscope (SEM). (D,I) Human neutrophils were treated with PMA or MDA-MB-231-TRAPs in the presence or absence of a PAD4 inhibitor (GSK484, 10 µM) or DNase I (10 U) for 3 hours. Mouse neutrophils received similar treatments, except that they were stimulated with 4T1-TRAPs. Representative immunofluorescence costaining images of DNA (blue), H3Cit (red), NE (green), or MPO (green) were used to assess NET formation. (E,J) Quantification of NETs was performed by detecting the content of dsDNA and MPO-DNA complexes in the post-treatment supernatants of cultured human and mouse neutrophils. Data (mean±SEM) represent at least three independent experiments (*p< 0.05; **p< 0.01; ***p< 0.001; ****p< 0.0001). FCM: Flow Cytometry.

TRAP treatment was able to induce the formation of NETs in human and mouse neutrophils in both a dose-dependent and time-dependent manner (online supplemental figure S2A–D). On stimulation with TRAPs for 3 hours, neutrophils formed abundant reticular structures typical of NETs which could be observed under scanning electron microscopy (SEM) (figure 2C and H). Flow cytometric analysis showed an increase in neutrophils double-positive for MPO and H3Cit after TRAP treatment (online supplemental figure S2E). IF staining revealed the extensive extracellular release of MPO, NE, and H3Cit from TRAP-treated neutrophils (figure 2D and I). Treatment with GSK484, a PAD4 inhibitor, and DNase I reduced neutrophil MPO and H3Cit IF induced by TRAPs, indicating a reduction in NET levels. Similarly, significant increases in the levels of cf-dsDNA and MPO-DNA complexes were observed in the supernatants of neutrophil cultures induced by TRAPs while the level of NET characteristic molecules decreased following the addition of PAD4 inhibitors and DNase I (figure 2E and J). IF staining of extracellular DNA with SYTOX green confirmed the increase of TRAPs-induced NETs (online supplemental video 1), which diminished after the addition of GSK484 and DNase I (online supplemental figure S2F). Taken together, these findings indicate that TRAPs can induce NET formation in both human and mouse neutrophils in vitro.

TRAPs induce NET formation via the HMGB1-TLR4-Myd88-ERK/p38 signaling pathway

Next, the mechanism underlying TRAP-induced NET formation was investigated. To identify the molecular components in TRAPs responsible for NET production, protease K digestion and ultrasonic treatment were employed. These treatments reduced TRAP-induced NET formation (figure 3A,B), suggesting that surface proteins, predominantly induce NET production. Previous studies by our team have shown that proteins such as HMGB1, HSP60, HSP70, and HSP90α are present on TRAP surfaces (online supplemental figure S3A).21 Blocking HMGB1, but not Hsp60, Hsp70, or Hsp90α, on TRAPs partially mitigated TRAP-induced NET formation (figure 3C). Additionally, NET levels were significantly reduced on mouse neutrophils when stimulated with TRAPs derived from HMGB1 knockdown (KD) 4T1 cells (figure 3D,E,online supplemental figure S3B).

Figure 3

Tumor cell-released autophagosomes (TRAPs) induce neutrophil extracellular trap (NET) formation via the HMGB1-TLR4-Myd88-ERK/p38 signaling pathway. (A,B) Mouse neutrophils were stimulated with protease K, or ultrasonically treated TRAPs for 3 hours. (A) NETs were examined for DNA (DAPI and SYTOX Green) by fluorescence microscopy, and (B) the levels of dsDNA and myeloperoxidase-DNA (MPO-DNA) complexes in culture supernatants were detected by ELISA. (C) TRAPs pretreated with or without α-HSP27, α-HSP60, α-HSP70, α-HSP90, or α-HMGB1 blocking antibodies, were used to stimulate mouse neutrophils for 3 hours, and then dsDNA and MPO-DNA complexes were detected in culture supernatants. (D,E) Mouse neutrophils were subjected to TRAPHmgb1 KD or TRAPHmgb1 NC treatment for 3 hours, (D) NETs were identified by DNA (DAPI and SYTOX Green) by fluorescence microscopy, (E) and the levels of dsDNA and MPO-DNA complexes were detected in culture supernatants. (F) Schematic overview of the mouse tail vein injection with NS, TRAPHmgb1 KD, or TRAPHmgb1 NC. (G) Levels of dsDNA and MPO-DNA complexes in the plasma of mice were detected. (H,I) Representative images of immunofluorescence staining for DNA (blue), MPO (green), and H3Cit (red) in lung tissues on day 10. (J,K) Neutrophils pretreated with described inhibitor, or solvent control (DMSO) solutions for 2 hours, were then stimulated with TRAPs for 3 hours. (J) NETs were examined for DNA (DAPI and SYTOX Green) by fluorescence microscopy, and (K) culture supernatants were collected to detect dsDNA and MPO-DNA complexes. (L) Neutrophils from control and TLR4-knockout mice were stimulated with TRAPs for 3 hours, and the levels of dsDNA and MPO-DNA complexes in the culture supernatants were detected. (M) Western blotting analyses of ERK and p38 protein phosphorylation in neutrophils treated with TRAPs at different time points. (N) Neutrophils pretreated with ERK and p38 pathway inhibitors for 1 hour were stimulated with TRAPs for 3 hours, and the levels of dsDNA and MPO-DNA complexes in the culture supernatants were determined. (O) Western blotting analysis of ERK and p38 protein phosphorylation in neutrophils treated with TRAPHmgb1 KD and TRAPHmgb1 NC. (P) Neutrophils pretreated with described inhibitors for 2 hours were then stimulated with TRAPs for 1 hour. Western blotting was used to detect the phosphorylation of ERK and P38 proteins. Data (mean±SEM) represent three independent experiments (*p< 0.05; **p< 0.01; ***p< 0.001; ****p< 0.0001; NS, not significant).

To explore the role of surface HMGB1 on TRAPs in NET formation and PMN establishment in vivo (figure 3F), experiments were conducted using TRAPs from HMGB1 KD-4T1 cells. Following tail vein injection of these TRAPs in mice, a marked decrease in NETs formation was observed in both plasma (figure 3G) and lung tissue (figure 3H,I), alongside a reduction in the proportion of lung tissue neutrophils (online supplemental figure S3C). Moreover, on day 10 postinjection, a partial restoration of lung T-cell function was noted (online supplemental figure S3D), indicating that TRAPs induce NET formation predominantly via surface HMGB1.

The specific receptors through which mouse neutrophils recognize TRAPs to induce NETs are not fully understood. TLRs, key innate immune cell recognition receptors, are crucial regulators of neutrophil functions. Neutrophils pretreated with inhibitors targeting TLR2, TLR4, and MyD88 displayed different levels of NET formation, indicating the involvement of TLR4 and MyD88 in this process (figure 3J,K). Notably, neutrophils from TLR4-deficient (TLR4−/−) mice stimulated by TRAPs failed to effectively induce NET formation (figure 3L).

Further investigation into the downstream signaling pathways revealed that TRAP treatment led to the phosphorylation of the ERK and p38 pathways in neutrophils (figure 3M). The addition of ERK and p38 inhibitors (SCH772984 and BIRB796, respectively), significantly reduced NET formation (figure 3N). Levels of ERK and P38 phosphorylation in neutrophils decreased following stimulation with TRAPs from HMGB1 KD-4T1 cells (figure 3O), and pretreatment with ERK or P38 inhibitors further reduced phosphorylation levels (figure 3P). These findings suggest that TRAPs may induce NET production through the HMGB1-TLR4-MyD88-ERK/p38 signaling axis.

TRAP-induced NETs inhibit T-cell function in a PD-L1-dependent manner in vitro

Neutrophils from both humans and mice were next subjected to transcriptomic sequencing after TRAP treatment. The sequencing results highlighted the significant enrichment of genes associated with the inhibitory regulation of immune processes (figure 4A, online supplemental figure S4A), with PD-L1 identified as the predominant gene in this category (figure 4B, online supplemental figure S4B). These findings were corroborated by real-time PCR (online supplemental figure S4C). The presence of PD-L1 in TRAP-induced NETs, both in vitro (figure 4C,D) and in vivo (figure 4E), was confirmed through flow cytometry, cellular IF, and tissue IF staining. To further investigate whether PD-L1 in NETs originates from TRAPs or from the neutrophils themselves, 4T1 cells with PD-L1 knocked down were engineered (online supplemental figure S4D). Flow cytometry and IF analyses demonstrated that reducing PD-L1 in TRAPs does not affect PD-L1 expression in neutrophils nor the PD-L1 levels within NETs (figure 4C,D).

Figure 4

Tumor cell-released autophagosom (TRAP)-induced NETs inhibit T-cell function in a PD-L1-dependent manner. (A) RNA sequencing revealed the functional enrichment of differentially expressed genes in human peripheral blood neutrophils treated with or without MDA-MB-231-TRAPs (10 µg/mL) for 3 hours. (B) The heat map illustrates the genes differentially expressed between TRAP-treated and untreated neutrophils. (C) H3Cit and PD-L1 of TRAPPD-L1 NC and TRAPPD-L1 KD -stimulated mouse neutrophils for 6 hour were detected by FCM. (D) DNA (blue), H3Cit (red), MPO (green), and PD-L1 (orange) of TRAPPD-L1 NC and TRAPPD-L1 KD -stimulated mouse neutrophils for 6 hours were detected by immunofluorescence staining. (E) Representative images of immunofluorescence staining for DNA (blue), Ly6G (green), H3Cit (red), and PD-L1 (orange) in the lung tissues of mice after the tail vein injection of TRAPs. (F) Typical scanning electron microscope images of cell-free NETs isolated from TRAP-stimulated mouse neutrophils. (G) SEM images of cell-free NETs and coincubated T-cells. (H–K) TRAP-induced NETs were purified and incubated with either αPD-L1, DNase I, or αPD-L1 plus DNase I, and then cocultured with mouse splenocytes in the presence of anti-CD3/CD28 for 48 hours. The IFN-γ and Ki67 percentage of CD4+ and CD8+ T cells was detected by FCM. Data (mean±SEM) represent at least three independent experiments (*p< 0.05; **p< 0.01; ***p< 0.001; ****p< 0.0001; NS, not significant). FCM: Flow Cytometry.

To explore the potential effects of TRAP-induced NETs on T-cell function via PD-L1, purified NETs (figure 4F) were incubated with T cells (figure 4G). Flow cytometry analysis revealed that NETs significantly inhibited IFN-γ secretion (figure 4H, I) and proliferative capacity of CD4+ and CD8+ T cells (figure 4J,K). Blocking NETs with an αPD-L1 antibody restored T-cell function, with further enhancement observed on the addition of DNase I. These findings suggest that PD-L1 plays a key role in the inhibition of T-cell function by NETs.

αPD-L1 combined with DNase I attenuate inhibition of T-cell function by TRAP-induced NETs

To examine whether TRAP-induced NETs contribute to the formation of pulmonary PMN immunosuppression, two mouse models were established, including a model in which TRAPs were injected via the tail vein and an in situ 4T1 tumor-bearing model. Both models underwent treatment with αPD-L1, DNase I, and the combination of the two (figure 5A and E).

Figure 5

αPD-L1 combined with DNase I can attenuate tumor cell-released autophagosome (TRAP)-induced neutrophil extracellular traps (NETs)-mediated immunosuppression against T cells in vivo. (A,E) An experimental schematic detailing the establishment of the mouse model. (B,F) Lung-infiltrating lymphocytes were stimulated with the protein transport inhibitor, and intracellular IFN-γ expression in CD4+ and CD8+ T cells was analyzed by FCM. (C,G) Levels of plasma dsDNA and MPO-DNA complexes in the five groups were measured on day 10. (D,H) Luc-4T1 cells were injected intravenously for lung colonization analysis, with representative images of pulmonary surface nodules displayed on day 35. Data (mean±SEM) represent three independent experiments (*p< 0.05; **p< 0.01; ***p< 0.001; ****p< 0.0001; NS, not significant). MPO, myeloperoxidase. FCM: Flow Cytometry.

In the TRAP tail vein injection model, lung T-cell function was examined. Flow cytometry analysis revealed that both αPD-L1 and DNase I treatments, respectively, restored the function of CD4+ and CD8+ T cells in terms of IFN-γ secretion (figure 5B) and proliferation (online supplemental figure S5A), with further restoration observed in response to combination treatment. No significant decrease in plasma NETs was observed on day 10 in the αPD-L1 treatment group. However, NET levels were significantly decreased in both DNase I treatment and the combination treatment (figure 5C). After tail vein injection of luc-4T1 cells, a decrease in pulmonary metastasis was observed following individual treatments with αPD-L1 and DNase I, with a further reduction noted under their combined therapy (figure 5D). Lung metastasis diminished in the αPD-L1, DNase I, and combination treatment, indicating that TRAP-induced NETs may facilitate lung metastasis by promoting an immunosuppressive premetastatic microenvironment.

Similar results were obtained from the in situ 4T1 tumor-bearing model (figure 5E). Lung tissue infiltrating T-cell function was improved after αPD-L1 and DNase I treatment, as evidenced by increased IFN-γ secretion (figure 5F) and proliferation (online supplemental figure S5B) for both CD4+ and CD8+ T cells, with the combination therapy showing further recovery. NET levels after αPD-L1 treatment remained unchanged, whereas DNase I and combination treatment significantly decreased NET levels (figure 5G). Spontaneous pulmonary metastasis was reduced in the αPD-L1 and DNase I treatment, with an even more pronounced reduction in the combination treatment (figure 5H). Taken together, these findings suggest that TRAP-induced NETs enhance lung metastasis by creating a premetastatic immunosuppressive microenvironment in the lungs. This process can be effectively mitigated by DNase I-mediated NETs degradation and αPD-L1 therapy.

Circulating levels of TRAPs and NETs are correlated with lung metastasis of patients with BC

Finally, plasma samples were collected from patients with BC and healthy donors (HDs) to explore correlations between peripheral blood TRAP levels and NET levels, NET formation, and lung metastasis. Compared with HDs, patients with BC showed a marked increase in circulating TRAP levels (figure 6A,online supplemental figure S6A). Peripheral blood TRAPs in HDs exhibited minimal EpCAM expression while in patients with BC, over 80% of circulating TRAPs were EpCAM+, signifying their tumor-derived nature (online supplemental figure S6B).

Figure 6

Circulating levels of tumor cell-released autophagosome (TRAP) and neutrophil extracellular trap (NET) are correlated with lung metastasis of patients with breast cancer (BC). (A,B) The number of TRAPs in healthy donors (HDs, n=42) and BC (n=105) patients with different stages of disease (A). Comparison of TRAPs between patients with BC without metastasis (n=78) and those with lung metastasis (n=27) (B). (C,D) HMGB1 mean fluorescence intensity (MFI) on TRAPs was compared between HDs and patients with BC at various stages I, and between patients with BC without metastasis and those with lung metastasis (D). (E,F) Plasma levels of MPO-DNA were compared between HDs and patients with BC at different stages I, and between patients with BC without metastasis and those with lung metastasis (F). (G,H) Correlations between peripheral blood TRAP levels and surface HMGB1 MFI (G) and MPO-DNA (H) in 105 patients with BC. (I–O) Receiver operating characteristic (ROC) curves for TRAPs (I), HMGB1 MFI on TRAPs (J), MPO-DNA (K), the combination of TRAPs and HMGB1 MFI on TRAPs (L), the combination of TRAPs and MPO-DNA (M), the combination of HMGB1 MFI on TRAPs and MPO-DNA (N), and the combination of all three (TRAPs, HMGB1 MFI on TRAPs, and MPO-DNA) (O). Data (mean±SEM) represent three independent experiments (*p< 0.05; **p< 0.01; ***p< 0.001; ****p< 0.0001; NS, not significant). MPO, myeloperoxidase.

Further analyses revealed that circulating TRAP levels in patients with BC with lung metastasis were significantly elevated relative to those in patients with non-metastasis (figure 6B). In comparison to HDs, patients with BC exhibited elevated HMGB1 levels on TRAPs with significantly higher levels in individuals with stage IV disease compared with stages I/II and III (figure 6C, online supplemental figure S6C). Serum HMGB1 levels in patients with lung metastasis were markedly higher than those in patients with BC without metastasis (figure 6D) and positively correlated with circulating TRAPs (figure 6G). Additionally, plasma MPO-DNA levels were elevated in patients with BC compared with HDs, with levels in stage IV patients significantly surpassing those in individuals with stages I/II and III disease (figure 6E). Patients with BCs with lung metastasis had significantly higher plasma MPO-DNA levels than those without metastasis (figure 6F), and these levels were positively correlated with circulating levels of TRAPs (figure 6H). Univariate logistic regression analysis identified circulating TRAPs, HMGB1 levels on TRAPs, and NETs as risk factors for lung metastasis of BC. ROC curve analyses revealed that the AUC for circulating TRAPs, HMGB1, and MPO-DNA were 0.81, 0.67, and 0.69, respectively (figure 6I–K). Multivariate logistic regression analysis indicated that the AUC for HMGB1 combined with circulating TRAPs and the combination of circulating TRAPs with MPO-DNA and HMGB1 were 0.82, 0.81, and 0.73, respectively, suggesting the synergistic performance of these biomarkers (figure 6L–N). The combined AUC for all three was 0.82 (figure 6O), suggesting that TRAPs, HMGB1 on TRAPs surface, and plasma MPO-DNA levels could serve as predictive biomarkers for lung metastasis of patients with BC.


NETosis, a form of inflammatory cell death, is capable of ensnaring bacteria, fungi, protozoa, and viruses.11 Many different factors including microorganisms, LPS, PMA, and activated platelets can significantly activate neutrophils and induce the formation of NETs. Recent studies have shown that NETs are involved in tumor progression and metastasis.27 In the TME, inflammatory cytokines released by cancer cells, such as IL-8, G-CSF, CXCL1, CXCL2, and CTSC,28–30 can induce NET formation. Previous studies have reported that in diffuse large B-cell lymphoma (DLBCL), DLBCL-derived IL-8 interacts with CXCR2 on neutrophils to form NETs through the Src, p38, and ERK signaling pathways.31 In this study, we identified a novel NET induction mechanism, in which we discovered that TRAPs were able to stimulate neutrophils to form NETs. Furthermore, we found that TRAPs induced NET formation by activating the TLR4-Myd88-p38/ERK axis through the DAMP molecule HMGB1 on the surface of neutrophils while blocking the HMGB1–TLR4 axis inhibits this inducible NET formation.

It has been reported that NE in NETs can enhance mitochondrial ATP production and promote primary tumor growth by stimulating TLR4 on tumor cells.32 NET-DNA has both structural and signaling functions. It physically captures circulating tumor cells, guiding their traversal through the vascular barrier into target organs, including the liver.33 Furthermore, NET-DNA has chemotactic properties that enhance tumor-cell proliferation, adhesion, and migration through the signaling protein CCDC25.15 It has been reported that in mouse models of liver metastasis, PD-L1 can be embedded within NETs, resulting in T-cell failure and dysfunction in the TME.34 Notably, consistent with the previous findings,34 we discovered that PD-L1 was present within TRAP-induced NETs and capable of negatively regulating T-cell function. When NETs were treated with DNase I or an αPD-L1 antibody, T-cell proliferation and IFN-γ secretion were partially restored, and this restoration was further enhanced by treatment with the combination of DNase I and αPD-L1. Therefore, it is clear that the immunosuppressive potential of NETs is not limited to their physical properties or their ability to capture immune cells, but also extends to their molecular components including PD-L1. Our study also revealed that circulating TRAPs were able to induce lung-infiltrating neutrophils to release PD-L1-embedded NETs, fostering an immunosuppressive PMN and promoting lung metastasis. The inhibition of PD-L1 in NETs restored T-cell activity and enhanced antitumor immune responses.

Our findings are further supported by clinical data from patients with BC. We found that the levels of circulating TRAPs and NETs in the plasma of patients with BC with lung metastasis were significantly higher than those of patients without metastasis. Additionally, patients with BC with lung metastasis had higher levels of circulating NETs, positively correlated with TRAP levels. We also found that the HMGB1 level on circulating TRAPs correlated with the level of NETs in the peripheral blood and with lung metastasis in patients with BC. Targeting HMGB1 is able to reduce the formation of NETs, thus reducing their ability to promote lung metastasis. These clinical data suggest TRAPs and NETs are potential biomarkers of metastatic disease that can be leveraged to gage the metastatic risk for patients with BC.

In summary, our study revealed that TRAPs induce the formation of NETs through the HMGB1-TLR4-Myd88-p38/ERK signaling pathway and that PD-L1 present within NETs inhibit T-cell function and promote BC pulmonary metastasis (figure 7). These findings suggest that the combined levels of TRAPs, HMGB1 on TRAPs, and NETs can serve as potential biomarkers for predicting BC lung metastasis, offering a novel strategy for early detection and treatment of lung metastasis of patients with BC.

Figure 7

Schematic showing that TRAPs induce the release of PD-L1-decorated NETs, which suppress T-cell function to promote breast cancer pulmonary metastasis. NETS, neutrophil extracellular traps; TRAPs, tumor cell-released autophagosomes.

Data availability statement

No data are available. Not applicable.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and received approval from the Ethics Committee Medical College of Qingdao University and was conducted in accordance with the QDU-HEC-2023106 agreement, ensuring adherence to ethical standards and research protocols. Participants gave informed consent to participate in the study before taking part.


We sincerely thank Dr ChunguangYan (Department of Microbiology and Immunology, School of Medicine, Southwest University) and Dr Wei Huang (Key Laboratory of Critical Medicine, Department of Critical Medicine, Zhongda Hospital, Southeast University) for their useful discussions and constructive comments on our manuscript. We thank LetPub ( for linguistic assistance and pre-submission expert review.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • XZ, CW, XW and NP contributed equally.

  • Contributors LW, XZ, CW, X-RW and NP designed and discussed this research. XZ, CW, and X-RW performed the experiments. X-TS, BC, SZ, Y-TW and JC provided experimental support. SZ and JC helped in the collection of samples from the patients. YW, FZ and HC helped in the collection of data. LW, XZ, CW, X-RW and NP helped in data analysis and figure preparation. LW, XZ, NP and HC helped in manuscript writing and gave final approval of the manuscript. All authors analyzed and discussed the data. All authors read and approved the final manuscript. LW acts as the guarantor.

  • Funding National Natural Science Foundation of China, Lixin Wang, Grant/Award Numbers: 31670918, 31370895 and 31970849; Shiya Zheng: Grant/Award Number: 82303959.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.