Lamininϒ1 and collagen IV form 3D cages of ECM for megakaryocytes, directly connected to the sinusoidal basement membrane

(A) Left panel. Schematic representation of the experimental workflow for 2D imaging of immunostained bone marrow cryosections from WT mice. Confocal imaging is performed on single ultrathin sections with an axial resolution of 250 µm. Right panel. Representative 2D images of a sinusoid-associated megakaryocyte immunostained for laminin ϒ1 (red), GPIbβ (white) and FABP4 (cyan). Cell nuclei are visualized with DAPI (blue) (from one out of three independent IF experiments).

(B) Left panel. Schematic representation of the experimental workflow for 3D analysis of whole-mount bone marrow preparations from WT mice. A stack of confocal images covering half the depth of the megakaryocyte is acquired and then z-projected to create a maximum projection image. Right panel. Representative maximal projection images of sinusoid-associated megakaryocyte immunostained for laminin ϒ1 (red), GPIbβ (white) and FABP4 (cyan) (from one out of three independent IF experiments).

(C) Representative maximal projection images of sinusoid-associated megakaryocyte immunostained for collagen IV (green) and GPIbβ (white). The inset image shows the FABP4 (cyan) and megakaryocyte (white) immunostaining.

(D) Bone marrow-isolated megakaryocyte maintaining an ECM cage. Left panel. Schematic of the experimental procedure used to isolate mouse bone marrow megakaryocytes. Right panel. Maximal projection 3D images showing the persistence of the ECM cage (collagen IV in green) around a freshly isolated megakaryocyte (GPIbβ in white).

*, sinusoid lumen ; arrowheads, basement membrane-cage connections; BM, bone marrow; bm, basement membrane ; MK, megakaryocyte; Bars, 10 µm.

Reduced laminin ϒ1 cage and megakaryocyte-sinusoid interactions in Lamα4-/- mouse bone marrow.

(A-B) Depletion of laminin α4 leads to a reduction in the laminin ϒ1 deposition, but not in collagen IV, in the cage around megakaryocytes and in the sinusoid basement membrane.

A. Representative maximal projection images showing the immunostaining of laminin γ1 (red) or collagen IV (green) in the Lamα4-/- compared to control mice. Two magnifications are shown for Lama4-/- mouse.

B. Quantification of laminin γ1 and collagen IV surface coverage per megakaryocyte and per basement membrane surface (laminin: 6<cage<17 and 15<bm<17; collagen IV: 6<cage<16 and 11<bm<21 expressed as a percentage, ****P>0.001 unpaired t-Test).

(C-E) Depletion of laminin α4 leads to a decrease in the sinusoid-associated megakaryocytes.

C. Representative maximal projections showing the immunostaining of laminin γ1 (red) and megakaryocytes (GPIbβ in white) in the bone marrow of Lamα4-/- and control mice.

D. Quantification of the total number of megakaryocytes per surface unit (s.u., 12,945 μm2, n=3 for each genotype, 87<n<92, P=0.9228, ns, Mann-Whitney test).

E. Quantification of the sinusoid-associated megakaryocytes (n=3, ****P>0.001 unpaired t-Test) in control (grey) and Lamα4-/- (dark) mice. Arrows point to megakaryocytes which are not associated with sinusoids (pMK).

bm, basement membrane; MK, megakaryocyte; pMK, MK in the parenchyma; Bar, 10 µm.

Integrin-mediated control of the 3D ECM cage around megakaryocytes

(A) Representative 2D images of Pf4cre bone marrow cryosections (250 nm) showing a sinusoid-associated megakaryocyte immunostained for β1 integrin (MAB1997 in yellow). Right: the boxed area is shown at a higher magnification.

(B) Activated β1 integrins form functional adhesion structures around megakaryocyte surfaces. Representative 2D images of Pf4cre bone marrow cryosections showing laminin (red, upper panels) or collagen IV (green, lower panels) staining around a sinusoid-associated megakaryocyte immunostained for activated β1 integrin (9EG7 in cyan) and GPIbβ (white). Right: Magnification of the boxed area showing co-localization of the ECM proteins and activated β1 integrin.

(C) Depletion of β1 and β3 integrins leads to a reduction in the laminin deposition on the surface of megakaryocytes. Representative 3D images showing a decrease in laminin deposition (red) on Itgb1-/-/Itgb3-/- megakaryocytes compared to Pf4cre.

(D) Quantification of laminin surface coverage per megakaryocyte (in %, 17<n<19 as indicated in the bars, ***P>0.001 unpaired t-Test), (Middle) expression profile of the laminin staining along straight-lines (25 µm long) visible as white lines in the confocal images, and (Right) quantification of mesh sizes (in µm, 14<n<16, **P<0.01 Mann-Whitney).

*, sinusoid lumen; MFI, mean fluorescence intensity; MK, megakaryocyte; pm, plasma membrane; Bars, 10 µm.

Integrins protects megakaryocytes from entering the bloodstream as whole cells

(A-C)) Higher proportion of intravasation events in mice lacking β1/β3 integrins.

A. Representative confocal images of Pf4cre and Itgb1-/-/Itgb3-/-whole-mount bone marrow immunostained for GPIbβ (white) and FABP4 (cyan).

B. Quantification of megakaryocyte intravasation and circulating megakaryocytes (3 mice minimum for each genotype, 213<n<397 for Pf4cre and Itgb1-/-/Itgb3-/-, ** P<0.001 Mann Whitney).

C. Quantification of the laminin γ1 deposition in the ECM cage in single knockout integrins and in GPVI knockout. Quantification of the intravasation events in single knock-out mice showing that both integrins are essential for the proper anchoring of megakaryocytes in their vascular niche.

(D -E) Intravital two-photon imaging of Itgb1-/-/Itgb3-/- mouse calvarial bone marrow

D. Tissues were stained with intravenously injected AF488-conjugated anti-GPIX antibody and rhodamin dextran. The white arrow indicates an intrasinusoidal Itgb1-/-/Itgb3-/-megakaryocyte, dotted lines illustrate the sinusoid wall and the values in the left corner show the time-lapses.

E. Quantification of circulating megakaryocytes, expressed as a percentage of the total number of megakaryocyte (from 3 independent experiments, 130<n<136, 0.0279 *P<0.1, Paired t-test).

(F-G) Large megakaryocyte nuclei detected in the pulmonary capillaries of Itgb1-/-/Itgb3-/- mice

F. Representative confocal images showing megakaryocytes’ nucleus (arrow, GPIbβ green, DAPI nucleus) within the pulmonary microvessels of Itgb1-/-/Itgb3-/- mice. Cyan dotted lines indicate the vessel wall.

G. Quantification of the intravascular megakaryocytes (from 5 independent experiments, 28<n<149, 0.0079 **P<0.01, Mann-Whitney).

(H) Two TEM image showing intravascular entire Itgb1-/-/Itgb3-/- megakaryocyte.

*, sinusoid wall; FG, fibrinogen; FN, fibronectin; FG, fibrinogen; n, nucleus; sMK, sinusoid-associated MK; pMK, MK in the parenchyma; PPT, proplatelets; Bars in A-G, 10 μm; Bar in E, 5 µm; Bar in G, 30µm.

Integrins promote megakaryocyte adhesion to the ECM components of the bone marrow

(A-B) Impaired adhesion and spreading of Itgb1-/-/Itgb3-/- megakaryocytes.

A. Representative SEM images depicting bone marrow-derived megakaryocytes adhering on laminin.

B. Spreading (hatched bars) and round (grey bars) megakaryocytes were counted following 3h incubation on laminin, fibronectin (FN), and fibrinogen (FG) (in %) (from 4-6 independent experiments).

(C-D) Microfluidic flow chamber to study megakaryocyte adhesion efficiency.

C. Representatives bright field images showing that upon flow application, Itgb1-/-/Itgb3-/- megakaryocytes detach from fibrillary fibronectin protein, while Pf4Cre MKs remain attached.

D. Quantification of the detachment of Pf4cre and Itgb1-/-/Itgb3-/- megakaryocytes on laminin, fibrillar fibronectin and fibrinogen (from 5 to 7 independent experiments,**P<0.01, ***P<0,001, Mann-Whitney).

(E-F) Reduced physical anchoring of Itgb1-/-/Itgb3-/- megakaryocytes to bone marrow.

E. Representatives bright field images of the ten femur bone marrow sections placed in an incubation chamber (left panel), of the box (center panel) and of the megakaryocytes released from the periphery of the explants (right panel).

F. Quantification of the number of Pf4cre and Itgb1-/-/Itgb3-/- megakaryocytes released from the explants following 3h (from 10 to 13 independent experiments, 594<n<1095 for Pf4cre and Itgb1-/-/Itgb3-/-, *P<0.05, unpaired t-test).

dotted lines, MK detachment; MK, megakaryocytes, FN, fibronectin; FG, fibrinogen; n, number of cells studied; Bars in A, 10 μm; Bars in B, 20 µm; Bars in B, 30 µm.

Maturation of the 3D ECM cage is correlated with maturation of the DMS in megakaryocytes

(A-B) MMP inhibition leads to a densification of the 3D ECM cage.

A. Representative 3D confocal images showing a significant increase in collagen IV deposition (green) on the megakaryocyte surface in treated mice treated with the intravenous cocktail of protease inhibitors (B + I).

B. (Left) Quantification of collagen IV fluorescence showed a shortening of collagen IV fibers in treated mice compared to that in control mice (from 3-5 independent experiments, 20<n<22 as indicated in the bars, ****P<0;001, Mann-Whitney), (Middle) Histograms of fluorescence intensity versus distance showed an increase in cross-linking with a reduction in pore size (white lines of 25µm length are visible in the confocal images), (Right) Reduction in mesh sizes in treated mice (from 3 independent experiments, 7<n<12, ***P>0.001, t-test).

(C-D) MMP inhibition affects megakaryocyte growth.

C. Representative confocal images from DMSO vs B+I treated mice immunostained for GPIbβ (white) and FABP4 (cyan), zoomed-in images showing the difference in megakaryocyte size between the two groups (arrows).

D. Quantification of the number of megakaryocytes per bone marrow area (194 x 194 µm) (from 3 independent experiments, 229<n<483 as indicated in the bars, ****P<0.0001, Mann-Whitney).

(E-F) MMP inhibition leads to an increased in the number of immature megakaryocytes.

E. TEM observation revealed the presence of numerous immature megakaryocytes (stage II) in treated mice as compared to fully mature megakaryocytes in control mice (stage III).

F. Quantification of the total number of megakaryocytes (145<n<169 as indicated in bars, **P<0.01, Mann-Whitney) and in the proportion of immature megakaryocytes (stage II) in the B + I group (from 3 independent experiments, **P < 0.05, ***P < 0.01, one way ANOVA with Tukey correction).

(G-H) MMP inhibition reduced release of megakaryocytes from the bone marrow explant.

G. Representatives bright field images of the megakaryocytes (arrows) released from the periphery of the control and treated explants.

H. Quantification of the number of megakaryocytes released following 3h and 6h (from 6 independent experiments, 369<n<783 for DMSO and B+I, *P<0.05, unpaired t-test).

B + I, batimastat + Ilomastat; sMK, sinusoid-associated MK; pMK, MK in the parenchyma; n, number of cells studied; Bars in A and C, 10 μm; Bars in E, 5 µm; Bar in G, 50 µm.

Integrin-mediated signaling and MMP proteolysis regulate the matrix remodeling and the adhesive properties of the 3D ECM cage, which control megakaryocyte maturation and intravasation at the bone marrow-blood interface.

Molecular cartography of the ECM around megakaryocytes.

(A) Immunofluorescence 2D analysis of ECM proteins. Collagen IV (a and g), laminin γ1 (h), fibronectin (b), fibrinogen (c), von Willebrand factor (d), and collagen I /III (e-f) around sinusoid-associated megakaryocytes. Bone marrow immunostained for GPIbβ(white), FABP4 (cyan), and DAPI (blue).

(B) Positive immunostaining of small stromal cells near the bone with the antibody against collagen III.

(C) Positive immunostaining of the bone and arterioles using two antibody against collagen I (AB765P and ab21286, respectively). Note that the megakaryocytes are not immunostained.

(D-E) Direct connections between the basement membrane and the ECM cage

D. Maximal projection showing the interface between the basement membrane (red), the sinusoid (blue) and megakaryocyte (white). The next image is a 3D reconstruction of this area.

E. Maximum projection of 3D laminin ϒ1 fibers linked to the basement membrane. The arrowheads indicate the connextion between the basement membrane and the cage.

(F-H) The spatial distribution of the ECM cage is highly restrictive to the sinusoid-associated megakaryocytes

F. Maximal projection of a large field of the 3D laminin ϒ1 organization. Megakaryocytes in contact with sinusoids (sMK) have a cage, while megakaryocytes in the parenchyma (pMK) were surrounded by sparse interstitial laminin ϒ1. Bone marrow immunostained for GPIbβ(white) and laminin ϒ1 (red).

G. Quantification of the number of megakaryocytes displaying the typical 3D cage as a function of cell association with sinusoids. Megakaryocytes with an ECM cage (grey bars) and without a cage (hatched bars) were counted (three complete 3D stacks from independent mice, 126 megakaryocytes counted in total as indicated in the bars, **P<0.01 Mann-Whitney).

H. Quantification of the pMK relative to the total MK population. Megakaryocytes in the stroma and in contact with sinusoids were counted on three complete 3D stacks from independent mice, ***P<0.0001, unpaired t-test). OK

(I-J) Correlation with the maturation stages of megakaryocytes.

I. Maximal projections of megakaryocytes are shown. Bone marrow immunostained for GPIbβ(white) and laminin ϒ1 (red).

J. Quantification of laminin γ1 immunostaining on megakaryocytes. Quantification per cell shows the presence of a 3D ECM cage at all stages of maturation (2-3 independent experiments, 14<n<22 as indicated in the bars, *P<0,05 Mann-Whitney). Quantification on megakaryocytes extending proplatelets and residual megakaryocytes was not possible due to sample size limitations.

(K) Distribution of fibronectin and fibrinogen around sinusoid-associated megakaryocytes. Maximal projection 3D images showing the absence of ECM connections with the basement membrane for fibronectin (magenta) and fibrinogen (yellow). Sinusoid endothelial cells were visualized with anti-FABP4 (cyan) and megakaryocytes were identified with anti-GPIbβ (white).

Arrowheads, peri-MK staining; Arrow, staining in adjacent small cells in (Ag-h); bm, basement membrane; MK, megakaryocyte; DMS, demarcation membrane system; PPT, proplatelet; *, vascular lumen; Bar, 10 µm.

Characterization of Lamα4-/- platelets and megakaryocytes

(A-C) Lamα4-/- platelet morphology and functions.

A. Normal size, discoid shape and ultrastructure of Lamα4-/- platelets. Bar: 0.5 µm.

B. Gating strategy for platelet function analysis. Flow cytometry was used to analyse the functions of Lamα4-/- platelets. Following activation with collagen-related peptide (CRP)(40 µg/ml) or CRP + TRAP (4 mM), the expression of P-selectin (CD62P, a marker of α-granule secretion) and P-selectin (JON-A PE, a marker of integrin αIIbβ3 activation) was analyzed.

C. Quantification of the percentage of positive platelets, Bars are the mean ± standard error of the mean (SEM) of 3 to 6 independent experiments. The results were statistically compared to those of control platelets (*P < 0.05; **P < 0.01; ***P < 0.001).

(D-E) Megakaryopoiesis in Lamα4-/- mice.

D. Quantification of the bone marrow megakaryocytes observed by electron microscopy, per surface unit (s.u., 12,945 μm2). ns, p=0,7177 unpaired t test.

E. Representative TEM images of Lamα4+/+ and Lamα4−/− megakaryocytes. Bars, 5 μm.

(F-G) Ex vivo capacity of of Lamα4-/- megakaryocytes to form proplatelets.

F. Representative image showing proplatelets forming-Lamα4−/− megakaryocytes in fresh bone marrow explants. Arrowhead pointes to proplatelets. Scale Bar, 100 µm.

G. Quantification of the percentage of Lamα4+/+ and Lamα4−/− megakaryocytes extending proplatelets following 3h and 6h h. Bars represent the mean ± SEM of three independent experiments (107<n<113 for Lamα4+/+ and Lamα4−/−, ns p>0.1, unpaired t-test).

α, alpha granules, δ, dense granules, PT, platelet territories, bm, basement membrane, MK, megakaryocyte*, sinusoid lumen. Bars, 10 µm.

Integrins control the structural properties of the ECM cages around megakaryocytes

(A) Representative 2D images of sinusoid-associated megakaryocyte immunostained for β3 integrin (LucA5 in yellow). Right panel: the boxed area is shown at a higher magnification.

(B) Itgb1-/-/Itgb3-/- megakaryocytes did not display enhanced degradation of gelatin matrix. Freshly isolated megakaryocytes from Pf4cre or Itgb1-/-/Itgb3-/- bone marrow were deposited on a fluorescent gelatin matrix (green) for 3h and labeled with phalloidin (adhesive structures in red). ECM degradation was virtually absent in Pf4cre and Itgb1-/-/Itgb3-/- megakaryocytes. Tumoral 4T1 cells, which trigger large dark digested areas (arrow), were used as a positive control.

(C-D) Laminin is present in similar amounts in Itgb1-/-/Itgb3-/- and Pf4cre megakaryocytes.

C. Maximal projection 3D images illustrating the presence or not of granules containing laminin (in red) and fibrinogen (in yellow, as negative control) in megakaryocytes (in white) from freshly isolated Pf4cre and Itgb1-/-/Itgb3-/- megakaryocytes. The dotted lines delineated megakaryocytes.

D. Quantification of the number of granules per megakaryocyte (5<MK<6, ****P>0.001, t-test).

(E-F) Normal laminin deposition in the sinusoid basement membrane in Itgb1-/-/Itgb3-/-

E. Representative 3D images of basement membrane immunostained for laminin ϒ1 (red) (from one out of three independent IF experiments).

F. Quantification of fluorescence intensity of laminin per basement membrane surface (expressed as a percentage).

(G-H) Fibronectin and fibrinogen failed to form a ECM cage around Pf4cre and Itgb1-/-/Itgb3-/- megakaryocytes

G. Decrease in the expression of fibrillar fibronectin around Itgb1-/-/Itgb3-/- megakaryocytes.

H. Mislocalization of fibrinogen in Itgb1-/-/Itgb3-/- megakaryocytes.

Top panels. 2D immunofluorescence of Pf4cre and Itgb1-/-/Itgb3-/- bone marrow cryosections using anti-fibronectin (magenta) and anti-fibrinogen (green) antibodies.

Lower panels ImmunogoldEM images showing the fibronectin and fibrinogen stainings with 10 nm gold particles. Without integrins, megakaryocytes are unable to effectively remodel fibronectin into fibers. Fibrinogen staining is located in the α granules of Pf4cre megakaryocytes and retained in the extracellular DMS space of Itgb1-/-/Itgb3-/- megakaryocytes.

α, alpha granules; arrowheads, peri-MK staining; bm, basement membrane; MK, megakaryocyte; pm, plasma membrane;

*, sinusoid lumen; Bar, 10 µm.

Characterization of Itgb1-/-/Itgb3-/- platelets and megakaryocytes

(A) Itgb1-/-/Itgb3-/- platelet functional properties assessed by flow cytometry.

The gating strategy was similar to that in Supplemental Figure 2B. Bar graphs showing that activated Itgb1-/-/Itgb3-/- platelets expressed P-selectin but not JON-A PE. This confirmed their deficiency in integrin β3 (n=3 mice per genotype).

(B-E) Megakaryocyte density, maturation stages and ploidy were normal in Itgb1-/-/Itgb3-/- bone marrow

B. Quantification of the total number of megakaryocyte, observed by electron microscopy (from 6-7 independent experiments, 120<n<153 as indicated in the bars).

C. Representative TEM images of stage III megakaryocytes. (i) The DMS is well-defined with the presence of cytoplasmic territories in control mice. (ii) In Itgb1-/-/Itgb3-/- mice, the DMS accumulates in packets without marking territories (yellow arrow). Bars: 5 µm.

D. Classification of the megakaryocytes according to their maturation stage (from 3-5 independent experiments, 99<n<544).

E. Representative ploidy histograms of Pf4cre and Itgb1-/-/Itgb3-/- megakaryocyte (one out of three independent experiments).

(F) Intact megakaryocytes were undetectable in the blood of Itgb1-/-/Itgb3-/- mice.

Gating strategy for megakaryocyte identification: CD45+TER119-cells were selected from events with high FSC and SSC parameter (excluding platelets), CD41bright and CD42c bright events are megakaryocytes. The gate positions were determined based on FMO controls and biological positive controls (mouse bone marrow MK enriched cell suspension and whole blood supplemented with MK enriched cell population) (data not shown). Representative plots from 3 separate acquisitions on Pf4cre and Itgb1-/-/Itgb3-/- mice.

Adhesive properties of Itgb1-/-/Itgb3-/- megakaryocytes

(A) Schematic representation of the experimental microfluidic equipment.

(B) Determination of the seeding time with capture efficiencies of over 80 % for Pf4cre megakaryocytes on laminin-, fibronectin- or fibrinogen-coated chambers.

Proteolysis inhibition do not affect platelet count and in vitro megakaryocyte maturation.

(A) Experimental setup. A combination of batimastat (30 mg/kg) and ilomastat (20 mg/kg) (B + I) or vehicle (DMSO) was injected daily for 7 days.

(B-C) In vivo MMP inhibition is not impact platelet count and function.

B. Platelet counts were similar in both groups.

C. Flow cytometry shows normal P-selectin and JON-A PE expression following activation with collagen-related peptide (CRP)(40 µg/ml) or CRP + TRAP (4 mM). Bars graphs represent the percentage of positive platelets, mean ± standard error of the mean (SEM). The results were statistically compared to those of control platelets (ns, multiple unpaired t tests).

(D) In vitro studies demonstrate that batimastat and ilomastat have no significant direct effect on cultured megakaryocytes. This is evident from their lack of impact on : 1) proliferation (as expressed by the number of megakaryocytes (MKs) per unit), 2) maturation (as expressed by MK size in µm); 3) the ability to extend proplatelets (as expressed by the percentage of MKs that extend proplatelets). Bar graphs illustrate the impact of two MMP inhibitor doses on days 3 and 4 of the culture process, displaying the mean and standard error of the mean (SEM).

(E) Representative 2D images of bone marrow cryosections demonstrating an increase staining of activated β1 integrin (9EG7 in green) around megakaryocytes in bone marrow treated with B + I in comparison to that treated with DMSO.

*, lumen of the sinusoid; D, day; MK, megakaryocyte; Bars: 10 µm.

Image data flow graph