Dissection of platelet and myeloid cell defects by conditional targeting of the beta3-integrin subunit

Abstract

The purpose of this work was to determine platelet and myeloid cell-specific requirements for beta3-containing integrins in hemostasis, bone resorption, and tumor growth. LoxP-flanked mice were generated to study the conditional deletion of beta3-integrin in platelets [knockout in platelets (KOP)] and myeloid cells [knockout in myeloid (KOM)]. Using the beta3KOP and beta3KOM strains of mice, we studied the role of beta3-integrin in hemostasis, bone resorption, and subcutaneous tumor growth. Tissue-specific deletion of platelet beta3-integrins in beta3KOP mice did not affect bone mass but resulted in a severe bleeding phenotype. No growth difference of tumor xenografts or in neoangiogenesis were found in beta3KOP mice, in contrast to the defects observed in germline beta3(-/-) mice. Conditional deletion of myeloid beta3-integrins in beta3KOM mice resulted in osteopetrosis but had no effect on hemostasis or mortality. Tumor growth in beta3KOM mice was increased and accompanied by decreased macrophage infiltration, without increase in blood vessel number. Platelet beta3-integrin deficiency was sufficient to disrupt hemostasis but had no effect on bone mass or tumor growth. Myeloid-specific beta3-integrin deletion was sufficient to perturb bone mass and enhance tumor growth due to reduced macrophage infiltration in the tumors. These results suggest that beta3-integrins have cell-specific roles in complex biological processes.-Morgan, E. A., Schneider, J. G., Baroni, T. E., Uluçkan, O., Heller, E., Hurchla, M. A., Deng, H., Floyd, D., Berdy, A., Prior, J. L., Piwnica-Worms, D., Teitelbaum, S. L., Ross, F. P., Weilbaecher, K. N. Dissection of platelet and myeloid cell defects by conditional targeting of the beta3-integrin subunit.

Figure 1.

Successful disruption of the β3-integrin gene selectively in mouse platelets or OCs. A) β3-Integrin protein expression was absent in β3KOP mice. Representative Western blots of total β3-integrin protein in platelet lysates from 4- to 6-wk-old WT (first lane) and β3KOP mice (second lane). Total β-actin protein as loading control. B) β3-Integrin protein expression was absent during OC differentiation in β3KOM mice. Representative Western blots of total β3-integrin protein in OC lysates from WT and β3KOM mice at d 5 and 6 of OC differentiation. Lysates from d 6 β3^−/− (lane 5) and β3^+/+ (last lane) OCs were used as a control. Total β-actin protein as loading control. C) Absence of β3-integrin mRNA in lung endothelial cells (ECs) of WT mice. Endpoint RT-PCR was performed after EC isolation from pooled lung tissue (n=5) after a CD31 magnetic bead-mediated selection procedure. GAPDH was used as a housekeeping control gene. TEK, EC-specific receptor tyrosine kinase (∼200 bp); CO, control (no RT); LA, ladder. D) Detection of CD31 (red) and neighboring β3-integrin (green) in blood vessels recruited into a matrigel plug 10 d after subcutaneous placement of matrigel mixed with VEGF and heparin in β3KOP, β3KOM, and WT mice. Visualization by immunofluorescence. Nuclei are stained with DAPI. E) Number of vessel sprouts from aortic rings explanted from WT, β3KOP, β3KOM, and β3^−/− mice and cultured in matrigel in the presence of VEGF. **P < 0.01 for β3^−/− vs. WT mice. Data are means ± se.

Figure 2.

Disruption of β3-integrin in platelets and megakaryocytes results in a bleeding phenotype. A) No differences observed in red blood cell (RBC), white blood cell (WBC), or platelet counts in β3KOP (solid bars) vs. TS-HET (hatched bars) and WT mice (open bars). B) Increased gastrointestinal bleeding events in β3KOP mice. Fecal occult blood tests were performed on feces samples from β3KOP and littermate controls and demonstrated a significantly higher number of positive tests in β3KOP mice. *P < 0.01 vs. controls; χ² test. C) Increased bleeding times in β3KOP mice. Each symbol represents bleeding time on a single mouse according to genotype. In each case. bleeding in β3KOP mice had to be stopped by the examiner when bleeding did not stop spontaneously. P < 0.001; ANOVA with appropriate post hoc test. D, E) Platelet aggregation was dysfunctional in β3KOP platelets. D) In vivo clot retraction was assessed using pooled PRP from β3KOP mice and HET control mice. Red blood cells were added to enhance the color contrast for the photograph. E) Platelet aggregation assays on a slide (SPAT) were performed using PRP. No significant platelet aggregation was observed in PRP isolated from β3KOP mice; each test performed ≥3 times/genotype.

Figure 3.

β3KOP mice have normal bone mass and osteoclast function. A) Expression of Pf4 in OC results in a reduction of β3-integrin mRNA in OCs from β3KOP mice compared to OC from WT mice. RT-PCR was performed using mRNA harvested from OCs of β3KOP and WT littermates at d 1–3 of OC differentiation. A ∼300 bp band shows abundance of β3-integrin message in WT OCs and a distinct level of β3-integrin expression in OCs from β3KOP mice. LA, marker; CO, negative control. GAPDH was used as a housekeeping control gene. B) β3-Integrin protein expression is reduced, but not absent, in β3KOP OCs at d 6 of OC differentiation compared to WT OCs. β3-Integrin protein from d 6 OC from β3^−/− mice served as negative control and β-actin as loading control. C) OCs from WT, HET, and β3KOP mice demonstrated equivalent differentiation from bone marrow macrophages at d 2, 4, and 6 of OC differentiation. Representative TRAP staining of cultured macrophages and in vitro differentiated OCs from β3KOP mice and controls. D–F) BMD (D; n=6/group), serum CTX levels (E; n=10/group), and trabecular bone volume (F; n≥6/group) were equivalent in 6-mo-old WT (open bars), HET (hatched bars), and β3KOP mice (solid bars). G) Representative histology of femurs analyzed for trabecular bone volume.

Figure 4.

Myeloid-targeted β3KOM mice have intact platelet function. A) No significant difference in red blood cell (RBC), white blood cell (WBC), and platelet counts was detected in blood from β3KOM (solid bars), TS-HET (hatched bars), and WT mice (open bars). B) There was no significant difference in fecal occult blood testing in β3KOM, TS-HET, and WT mice. C) No significant difference in bleeding times of β3KOM, HET, TS-HET, and WT mice was measured. Each symbol represents bleeding time on a single mouse of appropriate genotype. D) Platelet aggregation was equivalent in β3KOM and WT mice. Platelet aggregation assays on a slide (SPAT) were performed using PRP; each test performed ≥3 times/genotype.

Figure 5.

Myeloid-specific ablation of the β3-integrin gene recapitulates the osteopetrotic phenotype described in β3^−/− mice. A, B) β3KOM mice (solid bars) demonstrated increased BMD compared to WT mice (open bars) at 2 mo (A) and at 6 mo of age (B). C) Trabecular bone volume was significantly increased in β3KOM mice vs. controls. D) Representative histology of femurs analyzed for trabecular bone volume (measured in C). Note the pronounced increase in trabecular bone in the β3KOM sample. E) CTX was significantly decreased in the serum of β3KOM mice vs. WT mice. F) Macrophages harvested and cultured from β3KOM mice demonstrated markedly dysfunctional differentiation into OCs on d 2–6 of OC differentiation compared to WT macrophages, as shown by representative in vitro TRAP staining. n ≥ 6/group. *P < 0.05; **P < 0.01.

Figure 6.

Subcutaneous tumor growth is enhanced in β3KOM, but not β3KOP, mice. A) In vivo bioluminescence at d 5, 7, and 10 after subcutaneous injection of B16-FL cells as assessed by photon flux in contoured ROI was not different between β3KOP mice (solid line), HET mice (hatched line), and WT mice (pointed line), n ≥ 5 tumors/group. Data are presented on a log-scaled y axis. B) Tumor-associated angiogenesis, as measured by immunohistochemical staining for CD31-positive blood vessels in 5-μm paraffin-embedded flank tumor sections counterstained with H&E and visualized by DAB, was equivalent in β3KOP mice (solid bar), HET controls (hatched bar), and WT mice (open bar); n ≥ 5/group. Data are presented as means ± sd. C) Macrophage infiltration, as measured by immunohistochemical staining for F4/80 in 5-μm paraffin-embedded flank tumor sections counterstained with H&E was equivalent in β3KOP, HET controls, and WT mice; n ≥ 5 tumors/group. D) In vivo bioluminescence at d 5, 7, and 10 after subcutaneous injection of B16-FL cells as assessed by photon flux in contoured region of interest (ROI) demonstrated increased bioluminescence in tumors from β3KOM mice (solid line) vs. WT mice (open line) at d 10; n ≥ 5/group. E) No difference in tumor-associated angiogenesis between β3KOM vs. WT mice, as measured by immunohistochemical staining for CD31-positive blood vessels in 5-μm paraffin-embedded flank tumor sections counterstained with H&E and visualized by DAB; n ≥ 5/group. F) Macrophage infiltration, as measured by immunohistochemical staining, for F4/80 or Mac-3 in 5-μm paraffin-embedded flank tumor sections counterstained with H&E, was decreased in tumors from β3KOM vs. WT mice; n ≥ 5/group. **P < 0.01.