Time-dependent interactions of blood platelets and cancer cells, accompanied by extramedullary hematopoiesis, lead to increased platelet activation and reactivity in a mouse orthotopic model of breast cancer - implications for pulmonary and liver metastasis

Abstract

Aging has become a significant risk factor for several diseases, including breast cancer.Platelet activation and platelet-cancer cell aggregate fractions were found to increase with tumor progression in a mouse model of breast cancer. At advanced stages of tumor development, platelets from mice with breast cancer were hyperreactive to low agonist concentrations and hyporeactive to high ones. Platelet activation and reactivity were strongly associated with breast cancer metastasis in the lungs and extramedullary hematopoiesis in the liver. A greater fraction of platelet aggregates was observed in 4T1-injected mice at the advanced stages of breast cancer. In vitro, platelet activation was elevated after incubation with 4T1 cells, and thrombin-stimulated platelets formed aggregates with 4T1 cells. Neither GPIbα, nor GPIIb/IIIa blocking antibodies, were able to affect platelet-cancer cell aggregation in vitro.The primed circulating platelets became more sensitive to subthreshold stimuli at advanced stages of tumor development, and the formation of platelet-cancer cell aggregates increased with cancer progression. Our findings demonstrate that the age-associated progression of breast cancer cells is connected with increased platelet functioning, and that it can be manifested by the increased number of metastases and extramedullary hematopoiesis in a time-dependent-manner.

Keywords: breast cancer; metastasis; platelets.

Figure 1

Representative histopathological images of breast cancer metastases in lungs of mice injected with 4T1 cancer cells. Breast cancer metastasis was investigated by histological examination of the lungs isolated from mice during necropsy carried out in animals injected with saline (control) (A) or after 0 (B), 2 (C), 3 (D), 4 (E) and 5 (F) weeks from the injection of 4T1 cancer cells. Hematoxylin and eosin staining, magnification of 100X. Additional images under lower optical magnification (20x) present breast cancer metastasis in samples of lungs resected from mice at the second (G) and fifth (H) week of breast cancer development. Cancer metastases are marked with white arrows. More experimental details are given in the Materials and methods section.

Figure 2

Representative histological images of the extramedullary hematopoiesis foci in livers of mice injected with 4T1 cancer cells. Extramedullary hematopoiesis was diagnosed histologically at increasing time intervals; observations were made based on slides from livers isolated during necropsy in animals injected with saline (control) (A) or after 0 (B), 2 (C), 3 (D), 4 (E) and 5 (F) weeks since the injection of 4T1-cancer cells. Hematoxylin and eosin staining, magnification of 100X. Additional images under lower optical magnification (20x) present extramedullary hematopoiesis foci in samples of lungs resected from mice at the second (G) and fifth (H) week of breast cancer development. The representative foci under higher magnification show the progenitor hematopoietic cells next to mature granulocytes (I) (400x) and a megakaryocyte (J) (600x). Extramedullary hematopoiesis foci are marked with white arrows. More experimental details are given in the Materials and methods section.

Figure 3

Representative images of immunochemistry detection of the extramedullary hematopoiesis foci in liver of mice injected with 4T1 cancer cells. Extramedullary hematopoiesis was diagnosed by immunohistochemistry staining at time interval t₅ in slides from liver isolated during necropsy in mice injected with 4T1-cancer cells. The expressions of hematopoietic markers: CD117 (erythroid marker) (A, B), MPO (granulopoietic marker) (C, D) and FVIII (hematopoietic markers for megakaryocyte) (E, F) were detected. Additional hematoxylin staining was applied. Magnification of 100X (A, C) and 400X (B, D, E, F). Extramedullary hematopoiesis foci are marked with white arrows. More experimental details are given in the Materials and methods section.

Figure 4

Representative histopathological images of the extramedullary hematopoiesis foci in spleen of mice injected with 4T1 cancer cells. Extramedullary hematopoiesis was diagnosed histologically at time interval t₅ in slides from spleens isolated during necropsy in mice injected with saline (control) (A, C, E) or 4T1 cancer cells (B, D, F). Hematoxylin and eosin staining, magnification of 20X (A, B), 100X (C, D) and 600X (E, F). Extramedullary hematopoiesis foci are marked with white arrows. More experimental details are given in the Materials and methods section.

Figure 5

Activation of circulating platelets in mice injected with 4T1 cancer cells or saline. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of P-selectin (CD62P) (A), the active form of GPIIb/IIIa (B) and the binding of endogenous vWF (C) and endogenous fibrinogen (Fg) (D) on resting platelets were measured using flow cytometry in non-fixed ‘washed blood’ withdrawn immediately (t₀) or after 5 weeks (t₅) from the injection of mice with 4T1 cancer cells or saline. Results are expressed as the percent fraction of platelets positive for a given activation marker. More experimental details are given in the Materials and methods section. The statistical significance of differences, estimated with Kruskal-Wallis test followed by the post hoc Conover-Inman all-pairwise comparisons test, P-selectin_resting, P_1,α < 0.001, 4T1 t₅ > saline t₅; P_1,α < 0.001, 4T1 t₅ > 4T1 t₀; active form of GPIIb/IIIa_resting, P_1,α < 0.001, 4T1 t₅ > saline t₅; P_1,α < 0.001, 4T1 t₅ > 4T1 t₀; vWF_resting, P_1,α < 0.001, 4T1 t₅ > saline t₅; P_1,α < 0.001, 4T1 t₅ > 4T1 t₀; Fg_resting, P_1,α < 0.001, 4T1 t₅ > saline t₅; P_1,α < 0.001, 4T1 t₅ > 4T1 t₀.

Figure 6

Time-course of circulating platelet activation in mice injected with 4T1 cancer cells during 5-week breast cancer development. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of P-selectin (CD62P) (A), the active form of GPIIb/IIIa (B) and the binding of endogenous vWF (C) and endogenous fibrinogen (Fg) (D) on resting platelets were measured using flow cytometry in non-fixed ‘washed blood’ withdraw immediately (t₀, open boxes) or after 2 (t₂, light grey boxes), 3 (t₃, grey boxes), 4 (t₄, dark grey boxes) or 5 weeks (t₅, black boxes) from the injection of 4T1-cells. Results are expressed as the percent fraction of platelets positive for a given activation marker. More experimental details are given in the Materials and methods section. The statistical significance of differences, estimated with Kruskal-Wallis test followed by post hoc all-pairwise comparisons Conover-Inman or one-way ANOVA followed by Tukey’s multiple comparisons test, was: P-selectin_resting, P_1,α < 0.05, 4T1 t₅ > 4T1 t₄ > 4T1 t₃ > 4T1 t₂ > 4T1 t₀; active form of GPIIb/IIIa_resting, P_1,α < 0.01, 4T1 t₅ > 4T1 t₄, 4T1 t₃, 4T1 t₂, 4T1 t_0; 4T1 t₄ > 4T1 t₂, 4T1 t₀; vWF_resting, P_1,α < 0.05, 4T1 t₅ > 4T1 t₄ = 4T1 t₃ = 4T1 t₂ > 4T1 t₀; Fg_resting, P_1,α < 0.01, 4T1 t₅ > 4T1 t₂, 4T1 t₀; 4T1 t₄ > 4T1 t₂, 4T1 t₀.

Figure 7

Expressions/bindings of selected platelet surface membrane activation markers on ADP-activated blood platelets in mice injected with 4T1 cancer cells or saline. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of the active form of GPIIb/IIIa (A, B) and binding of endogenous fibrinogen (Fg) (C, D) on platelets stimulated with ADP (5 or 20 μM) were measured using flow cytometry in non-fixed ‘washed blood’ withdraw immediately (t₀) (A, C) or after 5 weeks (t₅) (B, D) from the injection of 4T1-cells or saline. Results are expressed as the percent fraction of platelets positive for a given activation marker. More experimental details are given in the Materials and methods section. The statistical significance of differences, estimated with Kruskal-Wallis test followed by post hoc Conover-Inman all-pairwise comparisons or two-way ANOVA followed by Tukey’s multiple comparisons test, was: active form of GPIIb/IIIa_ADP5μM, P_1,α < 0.01, 4T1 t₀ > saline t₀; P_1,α < 0.01, 4T1 t₅ < saline t₅; active form of GPIIb/IIIa_ADP20μM, P_1,α < 0.01, 4T1 t₀ > saline t₀; P_1,α < 0.01, 4T1 t₅ < saline t₅; Fg_ADP5μM, P_1,α < 0.01, 4T1 t₀ < saline t₀; P_1,α < 0.01, 4T1 t₅ < saline t₅; Fg_ADP20μM, P_1,α < 0.01, 4T1 t₀ < saline t₀; P_1,α < 0.01, 4T1 t₅ < saline t₅.

Figure 8

Expression/binding of selected platelet surface membrane activation markers on thrombin-activated blood platelets in mice injected with 4T1 cancer cells or saline. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of P-selectin (CD62P) (A, B), the active form of GPIIb/IIIa (C, D) and binding of endogenous vWF (E, F) and endogenous fibrinogen (Fg) (G, H) on platelets stimulated with thrombin (0.025 or 0.25 U/ml) were measured using flow cytometry in non-fixed ‘washed blood’ withdrawn immediately (t_0, A, C, E, G) or after 5 weeks (t₅, B, D, F, H) from the injection of 4T1 cells or saline. Results are expressed as the percent fraction of platelets positive for a given activation marker. More experimental details are given in the Materials and methods section. The statistical significance of differences, estimated with Kruskal-Wallis test followed by post hoc all-pairwise comparisons Conover-Inman or two-way ANOVA followed by Tukey’s multiple comparisons test, was: P-selectin _thr0.025U/ml, P_1,α < 0.05, 4T1 t₀ < saline t₀; P_1,α < 0.01, 4T1 t₅ > saline t₅; active form of GPIIb/IIIa _thr0.025U/ml, P_1,α < 0.01, 4T1 t₅ > saline t₅; vWF _thr0.025U/ml, P_1,α < 0.05, 4T1 t₀ > saline t₀; P_1,α < 0.01, 4T1 t₅ > saline t₅; vWF _thr0.25U/ml, P_1,α < 0.05, 4T1 t₀ > saline t₀; Fg_thr0.025U/ml, P_1,α < 0.05, 4T1 t₀ < saline t₀; P_1,α < 0.01, 4T1 t₅ > saline t₅.

Figure 9

In vitro response to increasing concentrations of ADP of whole blood platelets from mice with orthotopic model of breast cancer. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of the active form of GPIIb/IIIa (A, B) and the binding of endogenous fibrinogen (Fg) (C, D) on platelets stimulated with 5 μM ADP (A, C) and 20 μM ADP (B, D) were measured using flow cytometry in non-fixed ‘washed blood’ withdrawn immediately (t₀) or after 2 (t₂), 3 (t₃), 4 (t₄) or 5 weeks (t₅) from the injection of 4T1 cells. Results are expressed as the percentage fraction of a given activation marker-positive platelets. More experimental details are given in the Materials and methods section. Statistical significance of differences, estimated with Kruskal-Wallis test followed by post hoc all-pairwise comparisons Conover-Inman or one-way ANOVA followed by Tukey’s multiple comparisons test, was: active form of GPIIb/IIIa_ADP5μM,; P_1,α < 0.05, 4T1 t₅ < 4T1 t₄ = 4T1 t₃ > 4T1 t₂ = 4T1 t₀; active form of GPIIb/IIIa_ADP20μM, P_1,α < 0.01, 4T1 t₄ > 4T1 t₂; Fg_ADP5μM, P_1,α < 0.05, 4T1 t₅ = 4T1 t₄ > 4T1 t₃ = 4T1 t₂ = 4T1 t₀; Fg_ADP20μM, P_1,α < 0.05, 4T1 t₅ = 4T1 t₄ > 4T1 t₃ < 4T1 t₂ > 4T1 t_0.

Figure 10

In vitro response of whole blood platelets from mice with orthotopic model of breast cancer to increasing concentrations of thrombin. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of P-selectin (CD62P) (A), the active form of GPIIb/IIIa (B) and binding of endogenous vWF (C) and endogenous fibrinogen (Fg) (D) on platelets stimulated with low thrombin (0.025 U/ml) (A, C, E, G) or high thrombin (0.25 U/ml) (B, D, F, H) were measured using flow cytometry in non-fixed ‘washed blood’ withdrawn immediately (t₀) or after 2 (t₂), 3 (t₃), 4 (t₄) or 5 weeks (t₅) from the injection of 4T1 cells. Results are expressed as the percent fraction of platelets positive for a given activation marker. More experimental details are given in the Materials and methods section. The statistical significance of differences, estimated with Kruskal-Wallis test followed by the post hoc Conover-Inman all-pairwise comparisons or one-way ANOVA followed by Tukey’s multiple comparisons test, was: P-selectin _thr0.025U/ml, P_1,α < 0.05, 4T1 t₅ = 4T1 t₄ > 4T1 t₃ > 4T1 t₂ > 4T1 t₀; P-selectin _thr0.25U/ml, P_1,α < 0.001, 4T1 t₅ > 4T1 t₄ = 4T1 t₃ > 4T1 t₂ < 4T1 t₀; active form of GPIIb/IIIa _thr0.025U/ml, P1_,α < 0.05, 4T1 t₅ < 4T1 t₄ = 4T1 t₃ > 4T1 t₂ = 4T1 t₀; vWF _thr0.025U/ml, P_1,α < 0.001, 4T1 t₅ = 4T1 t₄ = 4T1 t₃ > 4T1 t₂ = 4T1 t₀; vWF _thr0.25U/ml, P_1,α < 0.01, 4T1 t₅ = 4T1 t₄ > 4T1 t₃ < 4T1 t₂ = 4T1 t₀; Fg_{thr0.025U/ml,} P_1,α < 0.001, 4T1 t₅ = 4T1 t₄ = 4T1 t₃ > 4T1 t₂ = 4T1 t₀; Fg_thr0.25U/ml, P_1,α < 0.05, 4T1 t₅ < 4T1 t₄ = 4T1 t₃ > 4T1 t₂ = 4T1 t₀.

Figure 11

The formation of platelet-cancer cells aggregates in mice injected with 4T1 cells or saline. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of CD24 or CD44 within the population of the CD41/61-gated cells (platelets) were measured using flow cytometry in non-fixed ‘washed blood’ drawn immediately (t₀) or after 5 weeks (t₅) from the injection of 4T1 cells (n = 8). Results are expressed as the percent fraction of CD24/CD41/61 or CD44/CD41/61-positive cells. For further experimental details – see Materials and methods section. The statistical significance of differences, estimated with Kruskal-Wallis test followed by post hoc Conover-Inman all-pairwise comparisons or one-way ANOVA followed by Tukey’s multiple comparisons test, was: CD24/CD41/61-positive cells (4T1-platelet aggregates): P_1,α < 0.001, 4T1 t₀ > saline t₀, 4T1 t₅ > saline t₅; CD44/CD41/61-positive cells (4T1-platelet aggregates): P_1,α < 0.01, 4T1 t₀ > saline t₀; P_1,α < 0.01, 4T1 t₅ > saline t₅.

Figure 12

Time-course of the formation of platelet-cancer cell aggregates during 5-week tumor progression in the mouse model of breast cancer. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of CD24 (A) or CD44 (B) within the population of the CD41/61-gated cells (platelets) were measured using flow cytometry in non-fixed ‘washed blood’ withdrawn immediately (t₀) or after 2 (t₂), 3 (t₃), 4 (t₄) or 5 weeks (t₅) from the injection of 4T1 cells. Results are expressed as the percent fraction of CD24/CD41/61 or CD44/CD41/61-positive cells. More experimental details are given in the Materials and methods section. Statistical significance of differences, estimated with Kruskal-Wallis test followed by post hoc Conover-Inman all-pairwise comparisons or one-way ANOVA followed by Tukey’s multiple comparisons test, was: CD24/CD41/61-positive cells (4T1-platelet aggregates), P_1,α < 0.001, 4T1 t₅ > 4T1 t₀; P_1,α < 0.01, 4T1 t₅ > 4T1 t₃, 4T1 t₅ > 4T1 t₂; P_1,α < 0.05, 4T1 t₅ > 4T1 t₄, 4T1 t₄ > 4T1 t₀; CD44/CD41/61-positive cells (4T1-platelet aggregates), P_1,α < 0.01, 4T1 t₅ > 4T1 t₄ > 4T1 t₃ > 4T1 t₀; P_1,α < 0.01, 4T1 t₅ > 4T1 t₄ > 4T1 t₂ > 4T1 t₀.

Figure 13

Representative dot plots and histograms showing gating strategy for the detection of platelet-cancer cells aggregates in mice injected with 4T1 cells or saline. The parent gate was set to the cancer cells based on FSC/SSC physical parameters (gate P1, in blue). Next, within this gated population, platelet-cancer cells aggregates were identified according to the surface presence of both CD41/61 (a unique antigen for platelets) (gate P2, in green) and CD24 (marker for cancer cells) (gate P3, corrected on the isotype binding, in red) in blood samples taken from mice five weeks after the injection with saline (A) or 4T1 breast cancer cells (B).

Figure 14

In vitro activation of whole blood platelets incubated with 4T1 cells. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). The expressions of P-selectin (CD62P) (A), the active form of GPIIb/IIIa complex (B) and the binding of endogenous vWF (C) and endogenous fibrinogen (Fg) (D) on platelets was measured using flow cytometry in non-fixed ‘washed blood’ incubated with 4T1-cancer cells or saline. Results are expressed as the percent fraction of the CD41/61-positive platelets. More experimental details are given in the Materials and methods section. The statistical significance of differences, estimated with the Mann-Whitney U-test, was: P-selectin, P_1,α < 0.001, 4T1 >saline; active form of GPIIb/IIIa, P_1,α < 0.001, 4T1 > saline; vWF, P_1,α < 0.01, 4T1 > saline; Fg, P_1,α < 0.05, 4T1 > saline.

Figure 15

The effect of the inhibition of GPIIb/IIIa complex and GPIb on the formation of platelet-4T1 aggregates in washed blood samples. Results are presented as median (horizontal line) and interquartile range (box) (n = 8). Aliquots of washed blood, preincubated with blocking antibodies anti-GPIIb/IIIa (n = 8), anti-GPIbα or saline, were mixed with CellTracker-labeled 4T1 cancer cells or saline. Results are expressed as the percent fraction of CellTracker/CD41/61-positive cells (4T1-platelet aggregates). More experimental details are given in the Materials and methods section. The statistical significance of the differences, estimated with the Kruskal-Wallis’ test and the post hoc multiple comparisons Conover-Inman test, was: P_1,α < 0.001, saline < 4T1_saline = 4T1_GPIb = 4T1_GPIIb/IIIa.