Platelet Angiopoietin-1 Protects Against Murine Models of Tumor Metastasis

Abstract

BACKGROUND:

In addition to their fundamental roles in preserving vascular integrity, platelets also contribute to tumor angiogenesis and metastasis. However, despite being a reservoir for angiogenic and metastatic cytokines, platelets also harbor negative regulators of tumor progression. Angpt1 (angiopoietin-1) is a cytokine essential for developmental angiogenesis that also protects against tumor cell metastasis through an undefined mechanism. Although activated platelets release Angpt1 from α-granules into circulation, the contributions of platelet Angpt1 to tumor growth, angiogenesis, and metastasis have not been investigated.

METHODS:

Using cytokine arrays and ELISAs, we first compared platelet Angpt1 levels in breast and melanoma mouse tumor models to tumor-free controls. We then assessed tumor growth and metastasis in mice lacking megakaryocyte and platelet Angpt1 (Angpt1^{Plt KO}). The spontaneous metastasis of mammary-injected tumor cells to the lungs was quantified using RT-PCR (reverse transcription-polymerase chain reaction). The lung colonization of intravenously injected tumor cells and tumor cell extravasation were determined using fluorescent microscopy and flow cytometry.

RESULTS:

Platelet Angpt1 is selectively upregulated in the PyMT (polyoma middle tumor antigen) breast cancer mouse model, and platelets are the principal source of Angpt1 in blood circulation. While primary tumor growth and angiogenesis were unaffected, Angpt1^{Plt KO} mice had both increased spontaneous lung metastasis and tumor cell lung colonization following mammary or intravenous injection, respectively. Although platelet Angpt1 did not affect initial tumor cell entrapment in the lungs, Angpt1^{Plt KO} mice had increased tumor cell retention and extravasation. Serum from Angpt1^{Plt KO} mice increased endothelial permeability and reduced VE (vascular endothelial)-cadherin expression at endothelial junctions compared with serum from control mice (Angpt1^WT).

CONCLUSIONS:

Platelets provide an intravascular source of Angpt1 that restrains tumor metastasis by preserving the lung microvasculature to limit tumor cell extravasation.

Highlights.

• Platelet Angpt1 (angiopoietin-1) is increased in the PyMT (polyoma middle tumor antigen) breast cancer mouse model.

• Platelets are the principal source of Angpt1 in blood circulation.

• Tumor growth and angiogenesis are not affected by platelet Angpt1.

• Platelet Angpt1 limits tumor cell metastasis by preventing tumor cell extravasation.

Platelets are small, anucleate cell fragments produced by megakaryocytes. During their activation, platelets secrete growth factors and cytokines from α-granules that play essential roles in maintaining vascular integrity. However, many of these same cytokines have also been associated with promoting tumor angiogenesis and metastasis.^1,2

Angpt1 (angiopoietin-1) is an angiogenic cytokine that binds Tie2 (also known as Tek receptor tyrosine kinase) on the surface of endothelial cells to induce receptor clustering and autophosphorylation.^3,4 Once phosphorylated, Tie2 initiates an Akt-driven signaling cascade that prevents apoptosis, promotes cell quiescence, and strengthens endothelial junctions.^5–7 During inflammatory conditions, levels of circulating Angpt2 (angiopoietin-2) increase and disrupt Angpt1 signaling by binding to Tie2.^8,9 Angpt2 binding is less efficient at inducing receptor clustering,^3,10 leading to reduced Tie2 autophosphorylation, and subsequently impairs both prosurvival signaling and vascular integrity.^8,11

Although mice lacking Angpt1 die in utero (embryonic day 12.5) with a dilated vasculature that lacks branching and supportive pericytes,^12,13 the contributions of Angpt1 to the adult vasculature can be assessed by either conditionally deleting Angpt1 later in development (eg, embryonic day 16.5), inducing Angpt1 overexpression, or intravenous administration of recombinant Angpt1 protein. These approaches show that Angpt1 supplementation protects against vascular leakage, including sepsis,¹⁴ acute lung injury,¹⁵ and acute kidney injury.¹⁶ In breast and colorectal neoplasms, Tie2 is likely expressed in tumor vasculature,^17,18 and elevated Angpt2 expression relative to Angpt1 is associated with an increased risk of metastasis and poor survival outcomes.^18,19 Although patient tumor cells rarely express Angpt1,¹⁷ tumors derived from Angpt1-overexpressing cell lines have reduced growth rates in mice, suggesting Angpt1 can limit tumor progression.^17,20 Despite this, global deletion of endogenous Angpt1 has no effect on primary tumor growth but limits tumor metastasis through an undetermined mechanism,²¹ warranting further investigation into the effects of native Angpt1 in malignancy.

Angpt1 is expressed by mesenchymal cells, pericytes, and vascular smooth muscle cells,^12,22,23 which under basal conditions produce a basolateral supply for endothelial cells.²⁴ However, platelets provide an intravascular source of Angpt1 in blood circulation that can be released from α-granules upon platelet activation.^22,25,26 It is, therefore, likely that the platelet Angpt1 reservoir plays a particularly important role in processes associated with platelet activation, including cancer.²⁷ Despite this, the contribution of platelet Angpt1 to tumor growth and metastasis has not been explored.

In this study, we demonstrate that platelet Angpt1 is selectively elevated in mouse models of breast cancer and limits tumor cell lung colonization and extravasation. Our findings demonstrate that, although widely recognized for their roles in promoting tumor cell metastasis, platelets are also equipped with negative tumor regulators and likely play a more nuanced role in tumor metastasis.

MATERIALS AND METHODS

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Animals

Angpt1^fl/fl (RRID:IMSR_JAX:028925) and PF4-Cre (RRID:IMSR_JAX:008535) mice on a C57BL/6 background were purchased from The Jackson Laboratories and subsequently crossed to generate Angpt1^{fl/fl, PF4-Cre} mice. Angpt1^fl/fl and Angpt1^{fl/fl, PF4-Cre} mice are hereinafter referred to as Angpt1^WT and Angpt1^{Plt KO}, respectively. Experimental groups were assigned depending on the genotype of Angpt1^WT and Angpt1^{Plt KO} mice, and sex as a biological variable was considered based on the nature of tumor generation, for example, mammary tumors were generated in female mice. To create an autochthonous tumor model in which mice lack platelet Angpt1, B6.FVB-Tg(MMTV [mouse mammary tumor virus]-PyVT)634Mul/LellJ (PyMT [polyoma middle tumor antigen]) transgenic mice expressing the polyoma middle T oncogene under the control of the mouse mammary tumor virus long terminal repeat^28,29 were purchased from The Jackson Laboratories (RRID:IMSR_JAX:022974) and crossed with Angpt^{Plt KO} mice to create Angpt1^{Plt KO} PyMT mice. Angpt1^WT PyMT littermates were used as control animals. PyMT mice on an FVB background FVB/N-Tg(MMTV-PyVT)634Mul/J were purchased from The Jackson Laboratories (RRID:IMSR_JAX:002374) and tumor-free littermates used as tumor-free controls. Thpo (thrombopoietin)-deficient mice were generously donated by Dr Ann Mullally. In all cases, mice were euthanized first by CO₂ asphyxiation, followed by either cervical dislocation or exsanguination. Inclusion or exclusion criteria were not applicable for animal studies. All animal experiments were performed in accordance with relevant guidelines and regulations and were approved by the Brigham and Women’s Hospital Institutional Animal Care and Use Committee (protocol No. 2019N000011).

Cell Culture

AT-3 (Sigma; No. SCC178, RRID:CVCL_VR89), MDA-MB-231 (ATCC [American Type Culture Collection]; HTB-26, RRID:CVCL_0062), B16-F10 (ATCC; CRL-6475, RRID:CVCL_0159), and E0771.lmb (ATCC; CRL-3405, RRID:CVCL_B0A2) cells were cultured in DMEM supplemented with 10% (v/v) fetal bovine serum and 1% (v/v) penicillin/streptomycin. Human primary umbilical vein endothelial cells (HUVECs) were purchased from ATCC (PCS-100-010, RRID:CVCL_2959) and cultured in Endothelial Cell Growth Media (R&D Systems; CCM027). All cell lines routinely tested negative for mycoplasma contamination.

Murine Platelet, Plasma, and Serum Isolation

For nonterminal murine blood collection, lateral incisions were made across a tail vein, and 15 μL of blood was collected using a pipette with an EDTA-prewetted tip. Blood was diluted 1:10 in EDTA (5 mmol/L), and full blood counts were assessed using the Sysmex Hematology XE-5000 analyzer. For terminal procedures, blood was collected via cardiac puncture into syringes containing EDTA (5 mmol/L) and transferred into EDTA-coated tubes (BD; 365974). Blood was spun (117g, 8 minutes) to separate platelet-rich plasma from red blood cells and white blood cells, with additional centrifugation (200g, 3 minutes) to purify platelet-rich plasma. For platelet isolation, platelet-rich plasma was supplemented with prostaglandin E₁ (1 μmol/L; Sigma-Aldrich; P5515), spun (1000g, 5 minutes), and resuspended in platelet-resuspension buffer (10 mmol/L HEPES, 140 mmol/L NaCl, 3 mmol/L KCl, 0.5 mmol/L MgCl₂, NaHCO₃, 10 mmol/L glucose, pH 7.4). To create platelet lysates, platelets were spun (1000g, 5 minutes) and lysed using the RIPA (radioimmunoprecipitation assay) Lysis Buffer System (Santa Cruz Biotechnology; sc-24948). Plasma was obtained by centrifugation of platelet-poor plasma at 10 000g for 5 minutes, and serum was collected from whole blood left at 37 °C for 30 minutes without an anticoagulant. Serum was purified by centrifugation (10 000g, 5 minutes).

Full Blood Counts

Murine blood samples were diluted 1:10 in 2 mmol/L EDTA (Boston BioProducts; BM-150) and immediately processed on the Sysmex XN1000V Hematology Analyzer to determine platelet counts, mean platelet volume, and the immature platelet fraction.

Flow Cytometric Analysis of Platelet Degranulation

Mice were bled up to 50 μL into EDTA-coated tubes, and blood was washed twice (700g, 5 minutes) using HEPES-buffered Tyrode’s solution without Ca²⁺. Afterward, washed blood was resuspended in Tyrode’s solution containing 2 mmol/L CaCl₂ and stimulated with either thrombin (Sigma; SRP6556), CRPXL (cross-linked collagen-related peptide; Cambcol Laboratories), ADP (Sigma; 01905), or the thromboxane agonist U46619 (Cayman Chemical; 16450) for 15 minutes. Samples were stained for CD62P (P-selectin; WUG 1-9-FITC [fluorescein]; Emfret Analytics) to determine the extent of platelet α-granule release. The reaction was stopped by adding 500 μL PBS. Mean fluorescence intensity was assessed by flow cytometry (Accuri C6 Plus; BD Biosciences).

Megakaryocyte Polyploidization

Tibias and femurs were dissected from euthanized mice and underwent centrifugation (2500g, 1 minute) to isolate bone marrow cells, which were passed through 100-µm strainers (Corning; 352360) and washed in PBS, before red blood cell lysis (BD Biosciences; 555899). Bone marrow cells were fixed and permeabilized with ice-cold methanol, then stained with Alexa Fluor 488–conjugated anti-CD41/61 antibodies (Emfret; M021-1). Propidium iodide (Thermo Fisher; P1304MP) was used to stain DNA, and samples were analyzed using the Cytek Aurora spectral cytometer.

ELISA and Cytokine Profiling

Protein concentrations of Angpt1 in murine platelet lysates, plasma, and serum were determined using ELISA kits (Aviva Systems Biology; OKBB00588). Proteome Profiler Mouse XL Cytokine array (R&D Systems; ARY028) was used to quantify cytokine content in plasma or platelet lysates as per the manufacturer’s instructions, with densitometry analysis conducted using ImageJ (ImageJ, RRID:SCR_003070). For cytokine arrays, 1 array was used per condition and consisted of 5 equally pooled platelet lysates or plasma samples.

SDS-PAGE Electrophoresis and Western Blot

Platelets, cell lines, and tissues were lysed with either the NP-40 buffer (Nonoxynol-40; Thermo Fisher Scientific; J62805.EQE) or RIPA buffer (Boston BioProducts; BP-115). Lysates were supplemented with Complete Protease Inhibitor Cocktail (Sigma; 11697498001) and stored at −80 °C. Lysates were reduced 1:1 in Laemmli Sample Buffer (Bio-Rad; 161-0737) and 1:1 NuPAGE Sample Reducing Agent (Invitrogen; NP0009), boiled (5 minutes at 95 °C), and loaded into a 4%-to-15% Mini-PROTEAN precast gel (Bio-Rad; 4561084). SDS-PAGE with Tris/glycine/SDS running buffer was performed, and the Bio-Rad Trans-Blot Turbo System was used to transfer proteins to nitrocellulose membranes (Bio-Rad; L002058). Membranes were blocked in SuperBlock T20 TBS (Thermo Fisher Scientific; 37536), before incubation with either Angpt1 (R&D Systems; catalog No. MAB8220, RRID:AB_3068348), Tie2 (Millipore Sigma; catalog No. 05-584, RRID:AB_11211769), or GAPDH (Cell Signaling Technologies; catalog No. 2118, RRID:AB_561053) primary antibodies. Samples were washed with TBS-T, then incubated with horseradish peroxidase–conjugated secondary antibodies (Invitrogen; catalog No. 65-6120, RRID:AB_88384; Jackson ImmunoResearch; catalog No. 712-035-150, RRID:AB_2340638). Membranes were treated with SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific; 34075), and chemiluminescence was measured using the ChemiDoc Imaging System (Bio-Rad).

Tumor Generation

For orthotopic mammary tumor generation, 2.5×10⁵ AT-3 cells were prepared in PBS and injected into the fourth left mammary fat pad of 9- to 13-week-old female Angpt1^WT and Angpt1^{Plt KO} mice. For melanoma tumor generation, 2.5×10⁵ B16-F10 cells in PBS were injected subcutaneously into 9- to 13-week-old male Angpt1^WT and Angpt1^{Plt KO} mice. Tumors were measured every 2 to 3 days using digital calipers, and tumor volume was calculated as (width²×length)/2. Mice were euthanized 28 and 16 days after AT-3 or B16-F10 injection, respectively. For spontaneous tumor generation, female Angpt1^WT PyMT and Angpt1^{Plt KO} PyMT mice were housed from birth and fat pads palpated from 8 weeks of age 2× to 3× per week for tumor incidence. Collective tumor volume measurements were made using digital calipers as described above.

Vascular Density

For immunofluorescent imaging of tumor vessels, tumors were fixed in 10% formalin solution (Sigma-Aldrich; HT5012) and stored in 70% ethanol before paraffin embedding and sectioning by iHisto, Inc. Antigen retrieval was performed using a citrate buffer (10 mmol/L citric acid, 0.05% Tween 20, pH 6.0), and sections were labeled first with antibodies against CD31 (R&D Systems; catalog No. AF3628, RRID:AB_2161028), followed by fluorescently conjugated donkey anti-goat secondary antibodies (Thermo Fisher Scientific; catalog No. A-21447, RRID:AB_2535864). CD31 surface coverage and signal intensity were quantified using the Gen5 software (BioTek, version 3.10).

Vascular Permeability by Evans Blue

Vascular permeability was assessed through the utilization of Evans blue dye (Thermo Scientific; A16774.09). Evans blue dye was administered intravenously (50 mg/kg) into anesthetized mice via the external jugular vein. After 20 minutes, organs were perfused with sodium citrate using an infusion pump (Harvard Apparatus, 22/2000). Lungs and tumors were then collected, weighed, homogenized in formamide (Thermo Scientific; 17899) for Evans blue dye extraction, and left at 55 °C overnight. Spectrophotometric quantification of the extracted dye was conducted at 620 nm using a microplate reader (Molecular Devices; SpectraMax ABS), with the degree of dye leakage indicative of vascular permeability.

Experimental Lung Metastasis

For breast tumor cell experimental metastasis, AT-3 cells were labeled with CellTracker Green CMFDA (5-chloromethylfluorescein diacetate; Thermo Fisher Scientific; C2925) and 2.5×10⁵ cells in PBS injected via the lateral tail vein of 9- to 12-week-old age-matched Angpt1^WT and Angpt1^{Plt KO} female mice. For flow cytometry analysis, lung tissue was digested with Collagenase/Dispase (Roche; 10269638001) and 100 μg/mL DNAase I (Roche; 11284932001) for 30 minutes at 37 °C. Tissues were dissociated using gentleMACS C Tubes (Miltenyi Biotec; 130-093-237) and samples then passed through 70-µm strainers before red blood cell lysis (BioLegend; No. 420301). Cells were resuspended in autoMACS Running Buffer (Miltenyi Biotec; 130-091-221) and CMDFA⁺ cells detected using the Cytek Aurora spectral cytometer. For immunofluorescent imaging, the left lung was inflated and embedded in Scigen Tissue-Plus OCT Compound (Fisher; No. 23-730-571). Lungs were flash frozen and stored at −80 °C. Embedded lungs were cut at 10 µm using a cryostat (Leica), and sections were blocked using 3% (v/v) goat serum. Nuclei were visualized using DAPI (4′,6-diamidino-2-phenylindole). Image acquisition was performed using either a confocal microscope (Zeiss LSM 880, ≈20 Å) or the Lionheart FX automated microscope. CMDFA⁺ AT-3 cells were quantified in lung sections using the Gen5 software (BioTek; version 3.10). To visually quantify the number of intravascular, transmigratory, and extravascular tumor cells, the spatial position of CMFDA⁺ AT-3 cells in relation to laminin-stained blood vessels was characterized in a blinded fashion. For melanoma cell experimental metastasis, 2.5×10⁵ B16-F10 cells in PBS were injected into the lateral tail vein of 10- to 15-week-old age-matched male Angpt1^WT and Angpt1^{Plt KO} mice. Upon euthanization, mice were perfused with 5 mmol/L EDTA and lung nodules manually counted in a blinded fashion.

Quantifying Extravasation by Flow Cytometry

Eighteen hours after the injection of CMFDA-labeled AT-3 cells via the lateral tail vein, mice were injected with 10 mg/kg biotin-NHS (N-hydroxysuccinimido; Sigma-Aldrich; 203118) 5 minutes before euthanization. Lungs were isolated and cells dissociated as described above. Lung cells were stained with APC (allophycocyanin)-conjugated streptavidin (BD Biosciences; 554067) to detect biotin on the tumor cell surface and extent of streptavidin signal on CMFDA⁺ cells used to assign the intravascular (streptavidin^high) or extravascular (streptavidin^low) status of tumor cells. Samples were analyzed using the Cytek Aurora spectral cytometer.

Endothelial Permeability

HUVECs (5×10⁴) were seeded atop 0.4-µm transwell inserts within a 96-well plate (Corning; No. 7369) and cultured for 48 hours. Transwell media was exchanged for various conditioned media for 9 hours, then carefully replaced with 1 mg/mL of 70-KDa FITC-dextran (Thermo Fisher; No. D1823) in FluoroBrite DMEM (Thermo Fisher; No. A1896701) for 30 minutes, with FluoroBrite alone in the receiver well. Fluorescence was quantified using the Agilent BioTek Synergy H1 Microplate Reader.

Endothelial Junctions

HUVECs (5×10⁴) were seeded within μ-Slide VI 0.4 chambers (IBIDI; No. 80606) and cultured for 48 hours. HUVECs were treated with various conditioned media for 9 hours, then fixed/permeabilized in ice-cold methanol, blocked in normal donkey serum, and stained with VE (vascular endothelial)-cadherin (R&D Systems; catalog No. AF938, RRID:AB_2161028) and DAPI. Four randomly selected fields of view were captured per chamber and quantified using the Lionheart FX automated microscope. Thresholding of VE-cadherin staining for junction area within an image was performed using the Gen5 software (BioTek, version 3.10).

Statistical Analysis

Statistical analyses and graphical representation were performed using R (v4.0.4; The R Foundation for Statistical Computing, Austria). Sample sizes were not based on power analyses but instead were estimated based on prior experiments. Biological replicates, sources of error, and P values are all reported in figures and figure legends. The distribution of data was tested for normality using the Shapiro-Wilk test, with P≤0.05 considered abnormal. For normally distributed data, statistical tests included unpaired 2-tailed t tests, with P≤0.05 considered statistically significant. For non-normal distributions, statistical tests included nonparametric Mann-Whitney U tests to compare 2 independent groups (Figures 1F, 2B, 2C, 2E, 2G, 3B, 3G, 4C, 4G, 6H, and 6J; Figures S1D, S1I, S3D, S6B, S6C, and S8C). In all cases, P≤0.05 was considered statistically significant. Kaplan-Meier plots using log-rank testing were used to compare tumor incidence (Figure 3E) and survival (Figure 3H).

Figure 1.

Increased platelet Angpt1 (angiopoietin-1) in mice with mammary tumors. A and B, Cytokine arrays of pooled platelet lysates isolated from 13-week-old PyMT (polyoma middle tumor antigen) mice compared with age-matched WT (wild-type) control mice. n=5 mice per array; 1 array per condition. C, ELISA-based Angpt1 quantification of platelets isolated from 13-week-old WT (n=11) and PyMT (n=17) mice. D and E, Cytokine arrays of plasma pooled 13-week-old PyMT mice compared with WT mice (n=5 mice per condition, 1 array per condition). F, ELISA-based quantification of Angpt1 from plasma and serum of 13-week-old WT (plasma, n=18; serum, n=6) and PyMT (plasma, n=24; serum, n=9) mice. G, Uniform Manifold Approximation and Projection (UMAP) of Angpt1 expression in CD41⁺ bone marrow cells isolated from a 13-week-old PyMT mouse and a WT mouse from a previously published and reanalyzed publicly available dataset published in the Harvard Dataverse.³⁰ G1, G2/M, MK1, MK2, and MK3 represent megakaryocyte (CD41^high) subsets. H, Differential gene expression of Angpt1 from PyMT-derived and WT-derived megakaryocytes. Statistical analysis was performed using 2-sided unpaired t tests (C) and Mann-Whitney U tests (F).

Figure 2.

Endogenous Angpt1 (angiopoietin-1) is not required for megakaryopoiesis, platelet production, or platelet function. A, Western blot for Angpt1 of platelet lysates from Angpt1^fl/fl (Angpt1^WT) and Angpt1^fl/fl PF4-Cre (Angpt1^{Plt KO}) mice. B and C, ELISA-based quantification of Angpt1 from platelet lysates (Angpt1^WT, n=14 mice; Angpt1^{Plt KO}, n=10 mice), plasma (Angpt1^WT, n=22 mice; Angpt1^{Plt KO}, n=13 mice), and serum (Angpt1^WT, n=20 mice; Angpt1^{Plt KO}, n=12 mice). D, Representative images of bone marrow cryosections staining megakaryocytes (CD41, cyan) and blood vessels (laminin, magenta) in the femurs of Angpt1^WT and Angpt1^{Plt KO} mice. Scale bars=100 μm. E, Quantification of megakaryocyte (MK) abundance and area in femoral cryosections (n=3 mice per condition). F, Ploidy analysis of native megakaryocytes isolated from the femurs and tibias of Angpt1^WT and Angpt1^{Plt KO} mice (n=3 mice per condition). G, Platelet counts, mean platelet volume (MPV), and immature platelet fraction (IPF) of Angpt1^WT and Angpt1^{Plt KO} mice (n=7 mice per condition). H, Baseline (control) and induced platelet surface CD62P (P-selectin) expression, following stimulation with ADP, U46619, thrombin, and CRPXL (cross-linked collagen-related peptide; n=3 mice per condition). Statistical analysis was performed using 2-sided unpaired t tests (E) and Mann-Whitney U tests (B, C, F, G, and H).

Figure 3.

Primary tumor growth is not dependent on platelet Angpt1 (angiopoietin-1). A, Tumor volume in Angpt1^fl/fl (Angpt1^WT, n=28) and Angpt1^fl/fl PF4-Cre (Angpt1^{Plt KO}, n=18) mice following mammary injection of AT-3 mammary tumor cells. B, Terminal tumor mass 26 days after the mammary injection of AT-3 cells into Angpt1^WT (n=25) and Angpt1^{Plt KO} (n=15) mice. C, Tumor volume in Angpt1^WT (n=8) and Angpt1^{Plt KO} (n=5) mice following subcutaneous injection of B16-F10 melanoma cells. D, Terminal tumor mass 16 days after the subcutaneous injection of B16-F10 cells into Angpt1^WT (n=8) and Angpt1^{Plt KO} (n=5) mice. E, The first recorded detection of mammary tumors in Angpt1^WT PyMT (polyoma middle tumor antigen; n=20 mice) and Angpt1^{Plt KO} PyMT (n=17 mice) mice. F, Cumulative tumor volume in Angpt1^WT PyMT (n=22 mice) and Angpt1^{Plt KO} PyMT (n=19 mice) mice. G, Individual tumor mass in the mammary fat pads of Angpt1^WT PyMT (n=210 tumors) and Angpt1^{Plt KO} PyMT (n=130 tumors) mice. H, Survival curve for Angpt1^WT PyMT (n=20) and Angpt1^{Plt KO} PyMT (n=17) mice. I, Representative staining of blood vessels (CD31, pink) in AT-3 mammary tumors from Angpt1^WT and Angpt1^{Plt KO} mice. J and K, Quantification of tumor vessel coverage and CD31 intensity in AT-3 tumors (n=4 mice per condition). L, Tumor vessel permeability assessed by Evan blue (EB) extravasation into the tissue (n=4 mice per condition). Statistical analysis was performed using 2-sided unpaired t tests (D, J, K, and L), Mann-Whitney U tests (B and G), and log-rank tests for Kaplan-Meier incidence and survival curves (E and H).

Figure 4.

Platelet Angpt1 (angiopoietin-1) protects against mouse models of lung metastasis. A, E0771.lmb cells were orthotopically injected into Angpt1^WT (n=6) and Angpt1^{Plt KO} (n=7–8) mice. Mice were euthanized after 21 days, with (B) terminal tumor mass recorded. C, RT-PCR (reverse transcription-polymerase chain reaction) quantifying the relative expression of mCherry signal in the lungs as a proportion of β-actin expression. D, B16-F10 melanoma cells were intravenously injected into Angpt1^WT (n=13) and Angpt1^{Plt KO} (n=10) mice. Mice were euthanized after 10 days and (E) the number of lung tumor nodules quantified. F, Representative H&E (hematoxylin and eosin) staining of lungs from Angpt1^WT and Angpt1^{Plt KO} mice 10 days after the injection of B16-F10 cells. G, Area of individual lung tumors 10 days after the injection of B16-F10 cells into Angpt1^WT (n=4 mice, 79 tumors) and Angpt1^{Plt KO} (n=4 mice, 173 tumors) mice. Statistical analysis was performed using 2-sided unpaired t tests (B and E) and Mann-Whitney U tests (C and G).

Figure 6.

Platelet Angpt1 (angiopoietin-1) prevents tumor cell extravasation and endothelial permeabilization. A, Representative images highlighting AT-3 cells (cyan), vasculature (laminin, gray), and platelets (CD41, magenta) within lung cryosections 3 hours after the intravenous injection of AT-3 cells into Angpt1^WT and Angpt1^{Plt KO} mice. Dashed yellow line highlights the tumor cell perimeter. Scale bars=10 µm. B, Example images of intravascular (Intra), transmigratory (Trans), and extravascular (Extra) AT-3 cells (cyan) in the lungs. Dashed yellow lines highlight the perimeter of tumor cells. C, Quantification of intravascular, transmigratory, and extravascular tumor cells in the lungs 3, 12, and 24 hours after intravenous injection into Angpt1^WT and Angpt1^{Plt KO} mice. D, CMFDA (5-chloromethylfluorescein diacetate)-labeled AT-3 cells were intravenously injected into Angpt1^WT and Angpt1^{Plt KO} mice. After 18 hours, biotin was injected intravenously 5 minutes before euthanization to label intravascular tumor cells. E, Representative flow cytometry plots comparing the degree of biotin labeling on tumor cells using APC (allophycocyanin)-conjugated streptavidin. F, Quantification of streptavidin-APC median fluorescence intensity (MFI) on AT-3 cells (Angpt1^WT, n=8 mice; Angpt1^{Plt KO}, n=10 mice). G, The proportion of cells considered extravascular (Angpt1^WT, n=8 mice; Angpt1^{Plt KO}, n=10 mice). H, Endothelial permeability following treatment with control media (media), 200 ng/mL TNFα (tumor necrosis factor-α), or 20% serum from Angpt1^WT or Angpt1^{Plt KO} mice (n=3 mice per condition, each with 4 technical repeats). Representative images (I) and quantification (J) of human primary umbilical vein endothelial cell VE (vascular endotheilal)-cadherin junctions following treatment with control media, 200 ng/mL TNFα, or 20% serum from Angpt1^WT or Angpt1^{Plt KO} mice. For control media and TNFα: n=3 technical repeats, each with 4 randomly selected images. For serum: n=6 mice per condition, each with 4 randomly selected images. Statistical analysis was performed using 2-sided unpaired t tests (F and G) and Mann-Whitney U tests (H and J). DAPI indicates 4′,6-diamidino-2-phenylindole; FITC, fluorescein; and FSC, forward scatter.

RESULTS

Increased Platelet Angpt1 in Mice With Mammary Tumors

Our first experiments assessed the systemic effects of breast cancer on platelet content by conducting an unbiased screen to identify cytokines selectively enriched in the platelets of tumor-bearing (PyMT) mice. Although cytokine arrays suggested increased platelet CXCL5 (C-X-C motif chemokine 5; LIX), CD62P, LCN2 (lipocalin-2), and OPN (osteopontin; Figure 1A and 1B), these factors were also upregulated in the plasma (Figure 1D and 1E) and were, therefore, not solely elevated in platelets. However, platelet Angpt1 was increased in both our cytokine array and follow-up ELISAs (Figure 1A through 1C), while plasma levels were decreased (Figure 1D through 1F). Serum, which includes the releasate of activated platelets, was unchanged in PyMT mice (Figure 1F), suggesting no differences in the collective (ie, intraplatelet and soluble) pool of circulating Angpt1. These findings collectively demonstrate that Angpt1 is selectively increased in the platelets of PyMT mice and can be released upon platelet activation. Platelet Angpt1 levels were unaffected in mice with subcutaneous melanoma tumors (Figure S1A through S1D), or dextran sulfate sodium-induced colitis (Figure S1E through S1K), demonstrating that platelet Angpt1 was only increased in the PyMT breast cancer model.

To test whether increased platelet Angpt1 in PyMT mice was due to differential gene expression by megakaryocytes, we utilized our previously published single-cell RNA-sequencing data set.³⁰ Angpt1 RNA transcripts were unaltered in megakaryocytes derived from a PyMT and WT (wild-type) mouse (Figure 1G and 1H), implying that changes in platelet Angpt1 levels occurred after platelet production. This theory was supported by experiments where platelet Angpt1 levels dose dependently increased following incubation with the recombinant Angpt1 protein, which can be partially blocked by targeting receptor-mediated endocytosis (Figure S2A and S2B). While Angpt1 levels increased in the platelets of PyMT mice, we found no difference in fibrinogen endocytosis in vitro (Figure S3), suggesting some selectivity for Angpt1 uptake into platelets.

Endogenous Angpt1 Is Not Required for Megakaryopoiesis, Platelet Production, or Platelet Function

Despite Angpt1 historically limiting both tumor growth and metastasis,^20,21 the contribution of platelet-derived Angpt1 to such processes has not been tested. We, therefore, created Angpt1^fl/fl PF4-Cre mice (Angpt1^{Plt KO}) selectively lacking Angpt1 in the megakaryocyte and platelet lineage, to compare against Angpt1^fl/fl (Angpt1^WT) controls (Figure 2A and 2B; Figure S4A through S4C). Plasma and serum levels of Angpt1 were ablated in Angpt1^{Plt KO} mice (Figure 2C), which matched the depletion of plasma Angpt1 in Thpo^−/− mice that have ≈10% of platelet counts (Figure S5A through S5D). We, therefore, concluded that platelets and megakaryocytes are the principal contributors to circulating Angpt1.

Before studying the effects of platelet Angpt1 on tumor biology, we first tested whether endogenous Angpt1 was required for megakaryopoiesis, platelet production, and platelet function. Angpt1^{Plt KO} mice had unaltered megakaryocyte size and numbers (Figure 2D and 2E), with typical ploidy distribution (Figure 2F), and no changes in platelet count, size, or maturity (Figure 2G; Figure S6A through S6C). These results are in line with previous work, demonstrating that Angpt1^{Plt KO} mice exhibit normal hematopoiesis and hematopoietic stem cell maintenance.³¹ Baseline and induced levels of platelet activation were also unchanged (Figure 2H). These findings establish that endogenous Angpt1 is not required for megakaryopoiesis, platelet production, or platelet activation.

Primary Tumor Growth Is Not Dependent on Platelet Angpt1

After determining that megakaryocytes and platelet Angpt1 were not required for platelet production or function, we next tested whether the absence of platelet Angpt1 affected primary tumor growth. Angpt1^{Plt KO} mice had no significant changes in tumor volume or terminal tumor mass in either breast cancer (AT-3 cells; Figure 3A and 3B) or melanoma (B16-F10 cells; Figure 3C and 3D) injection models. To test whether platelet Angpt1 was required for tumor growth in a spontaneous tumor model, we generated Angpt1^fl/fl PF4-Cre PyMT (Angpt1^{Plt KO} PyMT) mice, which developed palpable mammary tumors from ≈11 weeks of age (Figure 3E). Tumor incidence, cumulative tumor volume, terminal tumor mass, and survival were unchanged in Angpt1^{Plt KO} PyMT mice when compared with Angpt1^WT PyMT mice (Figure 3E through 3H), supporting our previous observations that platelet Angpt1 does not affect primary tumor growth.

Given that Angpt1 overexpression in tumor cells has paradoxically been shown to both increase³² and decrease³³ angiogenesis and vessel density within tumors, we sought to test whether removing platelet Angpt1 affected tumor neovasculature. Tumors from Angpt1^WT and Angpt1^{Plt KO} mice displayed no statistically significant difference in either blood vessel coverage or CD31 signal intensity (Figure 3I through 3K), and there was no difference in the extravasation of the azo dye Evans blue into the tumor tissue of Angpt1^WT and Angpt1^{Plt KO} mice (Figure 3L). Therefore, we conclude that deletion of platelet Angpt1 had no obvious effect on angiogenesis or vascular permeability within the primary tumor.

Platelet Angpt1 Protects Against Tumor Cell Experimental Metastasis

While deleting platelet Angpt1 did not influence the growth or vascularization of primary tumors, global Angpt1 has previously been implicated in reducing lung metastasis through an undetermined mechanism of action.²¹ To test whether platelet Angpt1 affected spontaneous lung metastasis, we orthotopically injected mCherry-expressing E0771.lmb cells into the left inguinal mammary fat pad of Angpt1^WT and Angpt1^{Plt KO} mice (Figure 4A). While there was no statistically significant increase in tumor growth or terminal tumor mass (Figure 4A and 4B), Angpt1^{Plt KO} mice had increased mCherry signal in the lungs (Figure 4C; Figure S7), indicative of increased metastatic burden. To test whether deleting platelet Angpt1 affected a genetically engineered mouse model, we also examined the lungs of Angpt1^WT PyMT and Angpt1^{Plt KO} PyMT mice but found no notable differences in the incidence or size of tumor nodules by hematoxylin and eosin staining and no clear difference in PyMT gene expression (Figure S8A through S8D). There are significant limitations of measuring lung metastasis with RT-PCR (reverse transcription-polymerase chain reaction),³⁴ and PyMT mice on a C57BL6/J background display both a high variability and low incidence of lung metastasis.^28,35 Therefore, we used synchronized blood-borne metastasis models to test whether removing platelet Angpt1 affected lung colonization by tumor cells and observed increased tumors in the lungs of Angpt1^{Plt KO} mice 10 days after the intravenous administration of B16-F10 cells (Figure 4D and 4E). Unlike mammary and subcutaneous tumors (Figure 3A through 3D), B16-F10 tumors in the lungs of Angpt1^{Plt KO} mice were larger than those within the lungs of Angpt1^WT mice (Figure 4F and 4G), suggesting that platelet Angpt1 selectively limits tumor outgrowth within the lung microenvironment. We also found increased AT-3 cells in the lungs of Angpt1^{Plt KO} mice 48 hours after intravenous injection (Figure 5A through 5D), demonstrating that platelet Angpt1 limits early lung colonization during the hematogenous stages of experimental metastasis. Unlike previously observed in PyMT mice (Figure 1C and 1E), the intravenous injection of B16-F10 or AT-3 cells did not influence Angpt1 expression in either platelets or plasma (Figure S9A through S9F).

Figure 5.

Platelet Angpt1 (angiopoietin-1) limits experimental lung metastasis after initial tumor cell seeding. A, CMFDA (5-chloromethylfluorescein diacetate)-labeled AT-3 tumor cells were intravenously injected into Angpt1^WT and Angpt1^{Plt KO} mice and euthanized after 48 hours. B, The number of AT-3 cells within the lungs was assessed by flow cytometry (n=6 mice per condition). C, Representative cryosections of lungs from Angpt1^WT (n=6) and Angpt1^{Plt KO} (n=5) mice 48 hours after intravenous tumor cell injection. Scale bars=100 µm; AT-3 cells in cyan. D, The number of AT-3 cells was quantified in relation to the lung area per section. E, CMFDA-labeled AT-3 tumor cells were intravenously injected into Angpt1^fl/fl (Angpt1^WT) and Angpt1^fl/fl PF4-Cre (Angpt1^{Plt KO}) mice; mice were euthanized after 3, 12, and 24 hours. F, representative plots and (G) quantification of the proportion of AT-3 cells in the lungs at different time points (Angpt1^WT: 3, 12, and 24 hours [n=6, 7, and 7 mice]; Angpt1^{Plt KO}: 3, 12, and 24 hours [n=7, 6, and 8 mice]). Statistical analysis was performed using 2-sided unpaired t tests (B, D, and G). DAPI indicates 4′,6-diamidino-2-phenylindole; and FSC, forward scatter.

To successfully metastasize, circulating tumor cells typically arrest within the blood vessels, survive immune- or shear-mediated clearance, and extravasate to colonize the extravascular space.³⁶ Because platelet Angpt1 deficiency increased experimental lung metastasis, we conducted time course experiments to test whether removing platelet Angpt1 promoted the initial arrest of tumor cells in the lungs or prevented their subsequent clearance (Figure 5E). We found similar numbers of tumor cells in the lungs of Angpt1^WT and Angpt1^{Plt KO} mice 3 hours after intravenous administration, signifying that platelet Angpt1 did not influence the initial tumor cell engraftment. However, tumor cell retention was greater in Angpt1^{Plt KO} lungs by 24 hours (Figure 5F and 5G), suggesting effects were only observed after the earliest hours of lung colonization.

Microscopic analysis of early experimental metastases highlighted platelet accumulation at sites of tumor cell arrest in the lungs of both Angpt1^WT and Angpt1^{Plt KO} mice (Figure 6A; Figure S10), theoretically providing a confined and concentrated intravascular source of Angpt1. Because Angpt1 mimetics can prevent tumor cell migration across an endothelial barrier,³⁷ we hypothesized that platelet-derived Angpt1 would limit tumor cell extravasation in the lungs. We found a higher proportion of extravascular tumor cells in Angpt1^{Plt KO} mice compared with Angpt1^WT controls by microscopy (Figure 6B and 6C) and flow cytometry (Figure 6D through 6G), demonstrating increased tumor cell extravasation in mice lacking platelet Angpt1.

Recently, docked platelets have been shown to release Angpt1, preventing vascular leakage during neutrophil extravasation.³⁸ Therefore, we hypothesized that platelet Angpt1 limits tumor cell extravasation by preserving the endothelial barrier. In vitro assays revealed that HUVECs treated with serum from Angpt1^{Plt KO} mice had increased permeability compared with Angpt1^WT controls (Figure 6H) and reduced VE-cadherin expression at cell-cell junctions (Figure 6I and 6J), demonstrating that platelet-derived Angpt1 can preserve integrity of the endothelial barrier and adherens junctions. To conclude, our data collectively show that platelet-derived Angpt1 preserves endothelial integrity and limits the successful extravasation of tumor cells.

DISCUSSION

In this article, we identify that platelet Angpt1 limits several mouse models of lung metastasis. Using multiple approaches, we demonstrate that while dispensable for primary tumor growth and angiogenesis, platelet Angpt1 reduces the spontaneous lung metastasis of tumor cells implanted into the mammary fat pad and prevents the lung colonization of melanoma and breast cancer cells following their injection into circulation. Mechanistically, platelet Angpt1 prevented tumor cell extravasation from lung vasculature, limiting both the retention and outgrowth of tumor cells within the lung microenvironment (Figure 7).

Figure 7.

Summary schematic: platelet Angpt1 (angiopoietin-1) limits lung colonization by preventing tumor cell extravasation. Under normal conditions where platelets contain Angpt1 (Angpt1^WT), circulating tumor cells arrest within lung vasculature and interact with platelets, leading to platelet degranulation and the release of Angpt1 into the vascular niche. Platelet Angpt1 prevents endothelial permeability and subsequent tumor cell extravasation, resulting in increased tumor cell clearance and reduced lung colonization. In situations lacking platelet Angpt1 (Angpt1^{Plt KO}), VE (vascular endothelial)-cadherin–based adherens junctions are lost, and endothelial permeability is increased at sites of tumor cell arrest, leading to tumor cell extravasation and proliferation within the lung microenvironment.

Platelets can sequester various angiogenic regulators, including fibroblast growth factor, endostatin, and vascular endothelial growth factor.³⁹ Because Angpt1 mRNA levels were unchanged in the megakaryocytes of PyMT mice compared with tumor-free controls and Angpt1 protein could be loaded into platelets in vitro, we hypothesize at least in PyMT mice that the platelet pool of Angpt1 increases due to uptake from the plasma pool into mature circulating platelets, which are capable of undergoing receptor-mediated endocytosis.^40,41 In contrast, the endocytosis of fibrinogen into WT and PyMT platelets was unchanged, implying a differential mechanism to be explored in future studies. We are also confident that varying platelet Angpt1 levels were not due to differences in platelet secretion, as PyMT mice have no abnormalities in platelet activation.³⁰ However, platelet Angpt1 levels were not increased after subcutaneous or intravascular injection of B16-F10 cells, or after intravascular injection of AT-3 cells, indicating that increased platelet Angpt1 is not a universal feature across tumor models. Interestingly, while platelet Angpt1 levels are increased in breast cancer patients, they are unchanged in patients with prostate cancer,⁴² highlighting that variable expression of platelet Angpt1 also extends beyond preclinical mouse models.

Despite the expression of Angpt1 in hematopoietic progenitors, stromal cells, pericytes, and smooth muscle cells,^12,23,31 platelet depletion reduces intratumoral levels of Angpt1,⁴³ and our data definitively show that megakaryocytes and platelets are the principal sources of Angpt1 in blood circulation. We, therefore, propose that platelet Angpt1 has particular importance in hemorrhagic and thrombotic regions where platelets accumulate. This hypothesis is supported by observations that platelet-depleted mice have increased intratumor hemorrhage, which is likely due to a lack of soluble mediators released by platelets during hemostasis, including Angpt1.⁴⁴ Similarly, in nonmalignant inflammatory models, platelet Angpt1 binding endothelial Tie2 has been implicated in preserving vascular integrity during neutrophil extravasation.³⁸

While several studies have established that the overall contribution of platelets is to promote metastasis,^45–47 our data reveal that platelets also contain negative metastatic regulators that can be released upon their activation. In addition to our data, platelet thrombospondin-1 is increased in tumor-bearing mice and limits the early stages of tumor angiogenesis,⁴⁸ suggesting that changes to the platelet proteome reflect a host response that attempts to suppress multiple aspects of tumor progression.

Although mice lacking platelet Angpt1 had increased spontaneous lung metastasis of E0771.lmb breast cancer cells implanted into the mammary fat pad, it should be noted that metastatic burden was unchanged in a PyMT model of platelet Angpt1 deficiency (Angpt1^{Plt KO} PyMT). This discrepancy likely stems from the high variability and lower incidence of the lung metastasis on a C57BL/6J background but should nonetheless be considered when discussing the translational impact of platelet Angpt1 in breast cancer metastasis.

Using experimental blood-borne metastasis models, we demonstrate that platelet Angpt1 limits tumor cell extravasation, which likely affects the retention and survival of tumor cells in the lung. Although deletion of platelet Angpt1 did not affect the growth of orthotopic mammary tumors or subcutaneous melanoma tumors, it surprisingly limited the size of experimental lung metastases. These findings suggest the protective effects of platelet Angpt1 extend beyond the early intravascular hours of lung colonization to also impair tumor outgrowth.

During inflammatory conditions, including cancer, circulating levels of Angpt2 increase due to the release of endothelial Weibel-Palade bodies and outcompete Angpt1 binding to Tie2, leading endothelial permeabilization and the extravasation of proangiogenic myeloid cells and cytokines.^8,49 This increase of plasma Angpt2 combined with absent platelet Angpt1 would likely synergize to promote inflammation, blood vessel permeability, and tumor cell metastasis. While endothelial cells abundantly express Tie2, both macrophages and pericytes can also express this receptor^50,51 and might also respond to platelet-derived Angpt1. However, Tie2 expression on macrophages appears dispensable for the formation of lung metastases,⁵⁰ and unlike endothelial cells, pericytes are not in direct contact with vascular sources of Angpt1. Our own data provide conflicting evidence for Tie2 expression in platelets (Figure S11), which was detectable by Western blot but not flow cytometry, implying some level of leukocyte contamination, and Tie2 was not detected in AT-3 tumor cells. These findings collectively suggest that platelet-derived Angpt1 exerts its effects via endothelial Tie2.

In conclusion, we demonstrate that platelet Angpt1 does not affect primary tumor growth or angiogenesis but instead limits the extravascular transition of tumor cells during lung colonization. This work provides a novel mechanism by which platelets can suppress tumor metastasis.

ARTICLE INFORMATION

Acknowledgments

H.G. Roweth proposed the research, designed experiments, analyzed data, and wrote the manuscript. H.G. Roweth, I.C. Becker, M.W. Malloy, E.M. Clarke, I. Aivasovsky, K. Tray, and A.A. Schmaier performed experiments. P.L. Kumar and S.A. Munn acquired images and analyzed data. A.A. Schmaier provided essential resources, including Angpt1^WT and Angpt1^{Plt KO} mice. I.C. Becker, A.A. Schmaier, and E.M. Battinelli provided feedback on the research and manuscript. Grants awarded to H.G. Roweth and E.M. Battinelli funded the research.

Sources of Funding

Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under award numbers K99CA283008 and 5R01CA200748. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work was also supported by the American Cancer Society (RSG-17-161-01).

Disclosures

None.

Supplemental Material

Supplemental Methods

Figures S1–S11

References

Supplementary Material

atv-44-2024-s001.pdf ^{(4.7MB, pdf)}