Programmable 3D silk bone marrow niche for platelet generation ex vivo and modeling of megakaryopoiesis pathologies

Abstract

We present a programmable bioengineered 3-dimensional silk-based bone marrow niche tissue system that successfully mimics the physiology of human bone marrow environment allowing us to manufacture functional human platelets ex vivo. Using stem/progenitor cells, megakaryocyte function and platelet generation were recorded in response to variations in extracellular matrix components, surface topography, stiffness, coculture with endothelial cells, and shear forces. Millions of human platelets were produced and showed to be functional based on multiple activation tests. Using adult hematopoietic progenitor cells our system demonstrated the ability to reproduce key steps of thrombopoiesis, including alterations observed in diseased states. A critical feature of the system is the use of natural silk protein biomaterial allowing us to leverage its biocompatibility, nonthrombogenic features, programmable mechanical properties, and surface binding of cytokines, extracellular matrix components, and endothelial-derived proteins. This in turn offers new opportunities for the study of blood component production ex vivo and provides a superior tissue system for the study of pathologic mechanisms of human platelet production.

Figure 1

Effect of silk film topography and stiffness on human megakaryocyte adhesion and proplatelet formation. (A) Silk films are prepared by dispensing a silk and PEO solution onto a PDMS mold. The surface of the mold may contain a grating pattern with defined depth and width. When the solution dries, a silk film is formed that contains a dispersion of PEO porogen. The film is finally soaked in phosphate-buffered saline to remove the PEO porogen. (B) Representative light microscopy image of silk film porosity (scale bar = 25 µm). (C-D) Analysis of Mk adhesion and proplatelet formation on silk film with different topography coated with fibrinogen (average ± standard deviatoin [SD], n = 3, *P < .05). Results are presented relative to silk film with no pattern. (E) Atomic force microscopy elastic modulus values obtained over hydrated low, medium, and high films. Distributions are displayed as percent of total sample points measured per bin. All samples had a minimum of 300 measurements. (F) There was no significant difference in Mk adhesion between the different stiffness samples (average ± SD, n = 4, P = not significant). (G) The low stiffness samples had similar proplatelet formation compared with the medium stiffness but significantly higher percentage compared with the high stiffness samples (average ± SD, n = 4, *P < .01). (H) Representative β1-tubulin staining of Mks cultured on silk films with different stiffness coated with fibrinogen, after a 16-hour incubation. The low stiffness silk films supported long proplatelet extensions and increased silk film stiffness appeared to decreased proplatelet branching (scale bar = 50 µm).

Figure 2

Effect of silk film functionalization on human megakaryocyte adhesion and proplatelet formation. (A-B) Mk adhesion and proplatelet formation on ECM entrapped within silk film followed a similar trend compared with ECM coated on glass coverslip or coated on silk film (average ± SD, n = 4, P = not significant). (C) Representative α-tubulin staining of Mks cultured on coated glass coverslip, coated silk film, or entrapped silk film. Mks were able to sense the proteins entrapped in silk film as they normally spread on type I collagen and form proplatelet on fibrinogen in all tested conditions (scale bar = 50 µm). (D-E) Analysis of silk film functionalization with bone marrow vascular niche ECM components: fibronectin (FNC), type IV collagen (COL IV), and laminin (LAM). Both Mk adhesion and proplatelet formation were not different between the 3 tested ECM components, but significantly higher compared with the nonfunctionalized silk film control only (average ± SD, n = 3, *P < .05). (F) Representative β1-tubulin staining of Mks cultured for 16 hours on functionalized silk films shows that proplatelet morphology was almost similar between the 3 tested conditions (scale bar = 50 µm).

Figure 3

Role of endothelial cells and endothelial-derived molecules in the regulation of platelet formation within the silk film culture system. (A) Schematic of the EPC and Mk seeding procedure for establishment of the silk film model. (B) After 16 hours of culture on the basal side of the silk film membrane, EPCs exhibited the characteristic cobblestone morphology and expression of VE-cadherin on both (Bi) glass coverslip control and (Bii) functionalized silk film (green = VE-cadherin, blue = nuclei, scale bar = 100 µm). (Biii-Biv) Representative fluorescent image of Mk and EPC coculture on the silk film culture system (green = VE-cadherin, red = CD61, blue = nuclei, scale bar = 50 µm). (Bv-Bvi) Representative cross-sectional image of Mk and EPC coculture rendered using confocal microscopy. There was distinct localization of the EPCs (green) on the basal side of the membrane and Mks (red) on the upper side of the membrane (green = VE-cadherin, red = CD61, blue = nuclear, scale bar = 20 µm). Silk films were stained with Hoechst 33258 and visualized in blue. (C) Analysis of Mk adhesion and proplatelet formation on silk film functionalized with ECM components in the presence or not of EPCs or VEGF and VCAM-1 (average ± SD, n = 6, *P < .05). (D) CD61⁺CD42b⁺ peripheral blood platelets were used to set the platelet gating protocol. Samples were mixed with counting beads to quantify the number of released platelets. Average ± SD of the mean fluorescence intensity of CD61 and CD42b staining from 6 different experiments is reported (P = not significant). (E) Mks cultured on functionalized silk film in the presence of EPCs or VEGF and VCAM-1 produced a significantly increased number of platelets compared with functionalized silk film only (average ± SD, n = 6, *P < .01).

Figure 4

Silk microtube fabrication and analysis of their ability to support platelet perfusion. (A) Silk microtubes are prepared by gel spinning aqueous silk solutions containing PEO porogen around a wire and functionalized via entrapment of ECM components. Resulting microtubes are freeze dried, removed from the wire, and soaked in water to leach out the PEO porogen. The resulting porous silk microtubes are fitted into the bioreactor chamber. (Ai) SEM cross sections of a silk microtube: microtube wall thickness was 50 ± 20 µm, with microtube wall pores diameter of 22 ± 4 μm to allow proplatelet elongation (scale bar = 20 µm). Arrows indicate silk microtubes borders. (Aii-Aiii) SEM images show pores on both the inner and outer surfaces of the silk microtubes, respectively. The inner and outer microtube wall pores diameter was 6 ± 2 μm (scale bars = 20 µm). (B) Whole blood (red) or peripheral blood platelets suspended in culture medium (pink) were perfused into functionalized silk microtubes. (C) Representative analysis of whole blood cells of 1 sample before (inlet) or after (outlet) perfusion. WBC, white blood cells; RBC, red blood cells; HCT, hematocrit; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; RDW, red blood cell distribution width; PLT, platelet. (D) Representative flow cytometry analysis of peripheral blood platelet basal activation before and after perfusion into silk microtube. Activation with ADP and thrombin demonstrated increased PAC-1 binding, indicating normal CD42b⁺ platelet functionality after the passage through the silk microtube lumen. (E) Confocal microscopy analysis of CD61⁺ platelet distribution within microtube lumen after passage of whole blood (green = CD61; blue = nuclei; scale bar = 50 μm). Silk fibroin microtubes were stained with Hoechst 33258 and visualized in blue.

Figure 5

Silk microtube assembly and sponge preparation into the bioreactor chamber and analysis of platelet production within this system. (A) Aqueous silk is dispensed into the chamber around the microtube and salt particles are added. After leaching out the salt, the resulting porous silk sponge is trimmed and sterilized. After seeding into the silk sponge, Mks migrate toward the microtube, adhere, and extend proplatelets through the microtube wall to release platelets into the microtube lumen. (Bi) SEM image showing a silk microtube embedded into the silk sponge (scale bar = 100 μm). (Bii) SEM image showing the porous morphology of silk sponge (scale bar = 100 μm). (C) Confocal microscopy analysis of different steps of Mks behavior within the silk microtube-sponge tissue system. (Ci) Silk sponge before Mks seeding (scale bar = 100 μm). (Cii) Mature Mks immediately after seeding into the silk sponge (green = CD61; blue = nuclei; scale bar = 100 μm). (Ciii) Migrated Mks in close contact with the microtube wall (green = CD61; blue = nuclei; scale bar = 50 μm). (Civ) Mks forming proplatelet on adhesion to the microtube wall (green = CD61; blue = nuclei; scale bar = 50 μm). (Cv) Mk extending proplatelets through the silk microtube wall. Arrow indicates proplatelet branch elongation through microtube wall with proplatelet tip protruding into the microtube lumen (green = CD61; blue = nuclei; scale bar = 50 μm). (Cvi) Boxes highlight proplatelet branches detectable along the inner wall of the silk microtube and platelets released into the microtube lumen (green = CD61; blue = nuclei; scale bar = 50 μm). For all immunofluorescence analysis, silk fibroin 3D scaffolds were stained with Hoechst 33258 and visualized in blue. (D) SEM imaging of different steps of Mks behavior within the silk microtube-sponge tissue system. (Di) Mature Mks adhesion on silk microtube outer wall (scale bar = 2 μm). (Dii-Diii) Migrated Mks forming proplatelet on adhesion to the microtube wall (scale bar = 2 and 10 μm). Arrows indicate silk pores that allow proplatelet elongation inside microtube lumen. (Div) Longitudinal section of silk microtube-sponge tissue system (scale bar = 200 μm). (Dv) Mk extending proplatelets through silk microtube wall (scale bar = 10 μm). (Dvi) Proplatelet branch stemming inside microtube lumen and released platelets (scale bar = 10 μm).

Figure 6

Analysis of ex vivo produced platelets morphology and functionality. (A) Silk microtubes are perfused with culture media for 6 hours, and released platelets are collected into gas-permeable bags. Samples are mixed with counting beads to quantify the number of platelets that are identified as CD61⁺CD42b⁺ events. A maximum of 4 different silk microtubes have been perfused concurrently. The graph shows the absolute number of platelets released per microtube embedded in the silk sponge containing 2.5 × 10⁵ Mks. (B) Analysis of platelet morphology. (Bi-Bii) Light microscopy analysis shows preplatelets, dumbbell-shaped platelets, and disc-shaped platelet (scale bar = 10 μm). (Biii) Immunofluorescence staining of β1-tubulin (green) (scale bar = 10 μm). (Biv-Bv) Magnification highlights the microtubule coil typically showed by resting platelets (scale bar = 5 μm). (Bvi) Transmission electron microscopy analysis of ex vivo produced platelets ultrastructure (scale bar = 2 μm). (C) Analysis of platelet adhesion on type I collagen. (Ci) Tetramethylrhodamine isothiocyanate-phalloidin staining of resting platelets (scale bar = 5 μm). (Cii) After adhesion on type I collagen platelets spread and formed filopodia/lamellipodia (scale bar = 5 μm). (Ciii-Civ) Magnification highlights actin cytoskeleton reorganization with evident stress fibers assembly (scale bar = 5 μm). (Cv) CFSE⁺ platelets were suspended into Tyrode’s buffer containing von Willebrand factor and perfused over immobilized type I collagen at shear rate of 1000 s^–1. Image shows a representative filed demonstrating platelet adhesion. Arrows indicate formation of platelet aggregates (scale bar = 10 μm). (D) Aggregation capacity was further measured by flow cytometry after stimulation with thrombin, ADP, and epinephrine. Platelets were separately labeled with CD31 or CD42b (left and right top, respectively), and then mixed 1:1, before being stimulated (bottom right) or not (bottom left) with the cocktail of agonists. (E) Analysis of ex vivo produced platelets participation to clot formation. (Ei) Ex vivo produced CFSE⁺ platelets were mixed with peripheral blood platelet stained with CellTracker Deep Red Dye. Clot formation was favored by addition of thrombin and visualized by confocal microscopy (scale bar = 10 μm). (Eii) Cross-sectional analysis of z-stack of the clot demonstrating that ex vivo produced platelets (green) actively interacted with in vivo derived platelets (red) with appearance of a juxtaposed signal (yellow) (scale bar = 5 μm).

Figure 7

Adding complexity to the system: endothelial and endothelial-derived molecules promote platelet collection. (A) The silk microtube lumen supports a confluent monolayer of human endothelial cells. (Bi) Confocal microscopy images of confluent HMVEC-d into the silk microtube lumen (green = VE-cadherin; scale bar = 100 µM). (Bii) Magnification of HMVEC-d seeded into silk microtube lumen (green = VE-cadherin; blue = nuclei; scale bar = 50 µM). (C) Statistical analysis of collected platelets after perfusion of silk microtubes in presence of endothelium or functionalized with VEGF and VCAM-1 with respect to silk microtube functionalized with ECM components only (average ± SD, n = 5, *P < .05).

Figure 8

Analysis of ex vivo platelet release by Mks differentiated from adult human peripheral blood hematopoietic progenitors. (A) Confocal microscopy imaging of Mks from 1 healthy control and 2 patients within the silk-based bone marrow system. (Ai-Aiii) Imaging of mature Mks immediately after seeding into the silk sponge (green = CD61; blue = nuclei; scale bar = 50 μm). (Aiv-Avi) Mks forming proplatelet through the silk microtube wall (green = CD61; blue = nuclei; scale bar = 50 μm). Arrows indicate proplatelet branching and elongation. Silk fibroin 3D scaffolds were stained with Hoechst 33258 and visualized in blue. (B) Flow cytometry analysis of ex vivo-produced platelets. Samples were mixed with counting beads to quantify the number of CD61⁺CD42b⁺ platelets. (C) Comparison of platelet count between in vivo and ex vivo quantified numbers.