Fabrication of Centimeter-Sized 3D Constructs with Patterned Endothelial Cells through Assembly of Cell-Laden Microbeads as a Potential Bone Graft

Abstract

Modular tissue engineering is a promising biofabrication strategy to create engineered bone grafts in a bottom-up manner, in which cell-laden micro-modules are prepared as basic building blocks to assemble macroscopic tissues via different integrating mechanisms. In this study, we prepared collagen microbeads loaded with human bone marrow derived mesenchymal stem cells (BMSCs) using a microfluidic approach. The cell-laden microbeads were characterized for size change, cell activity, osteogenesis, as well as their self-assembly properties to generate centimeter-sized constructs. Moreover, using the cell-laden beads as a supporting medium, induced pluripotent stem cell-derived endothelial cells (iPSC-EC) were patterned inside bead aggregates through extrusion-based 3D printing. This fabrication approach that combines modular tissue engineering and supports 3D printing has the potential to create 3D engineered bone grafts with a pre-existing, customized vasculature.

1. Introduction

Bone tissue plays critical roles in our body such as protecting various organs, sustaining hematopoiesis, storage of minerals, and providing mechanical support that enables mobility. Severe trauma, aging-related disorders, and diseases can lead to critical-size bone defects[1, 2], which require surgical bone implantation to restore structural integrity and functionality. Clinical bone repair imposes an enormous socio-economic burden: In the United States alone, approximately 1.6 million bone engraftments are performed on an annual basis, representing a cost of over $27 billion[3]. To date, bone autografts and allografts are considered as standard of care for critical bone defects. Nevertheless, these approaches are restricted to potential drawbacks including limited availability, donor site morbidity and pain, high failure rate, and risk of disease transmission[3–5]. The advancement of tissue engineering strategies, which combines cells, biocompatible scaffolds, and regenerative agents, presents marked advantages over the current bone graft sources[6, 7]. Tissue-engineered bone grafts (TEBGs) are generally composed of a biocompatible scaffold and autologous therapeutic cells such as bone marrow mesenchymal stem cells (BMSCs) that can populate the scaffold, elicit a regenerative response from host cells, and undergo osteogenesis to form the bone[5, 8–10]. These factors, along with selective growth factor components of the TEBG work synergistically to stimulate bone formation at the site of injury. Moreover, a successful TEBG also opens up unprecedented opportunities to study bone development and pathology ex vivo, as well as facilitate high-throughput drug screening[11].

Despite recent accomplishments in engineering tissue/organ replacements, creating large complex tissues in clinical relevant sizes remains extremely challenging due to the mass transfer limit in engineered tissues via diffusion, and limited success in hierarchical vascularization[12, 13]. Currently, modular tissue engineering is gaining increasing research interest as a new strategy of biofabrication in a bottom-up manner, in which cell-laden modules (microtissues) are prepared as basic building blocks to assemble macroscopic tissues (macrotissues) via different integrating mechanisms[14, 15]. The cell-laden modules in the size of several hundred microns are diffusible to external oxygen and nutrients, and thereby are highly viable and responsive to exogenous signaling cues[16]. Modular tissue engineering enables precise control over cellular spatial distribution in macroscale and offer great flexibility in fitting the clinical needs of customizable volume and architecture of grafts[17–19]. These features highlight the potential of modular tissue engineering as a promising avenue to create more advanced engineered bone grafts.

Sufficient vascularization of the engineered construct is the key to successful TEBGs due to several reasons. First, vascularization is essential to orchestrate the natural bone tissue regeneration. Many studies have indicated that angiogenesis precedes bone formation in various bone fracture models, and insufficient vascularity is associated with decreased bone formation[20, 21]. Second, the engineered vascular network greatly improves oxygen and nutrient availability to the cells seeded within scaffolds[22].

Third, the built-in vasculature enables rapid integration to the surrounding host vasculature via surgical or natural anastomosis to circumvent post-implant necrosis caused by hypoxia or starvation of the grafts. 3D bioprinting technology can deposit a great variety of cells, biomaterials, and regenerative agents in a layer-by-layer fashion with precise spatial control. It enables the fabrication of complex 3D tissue-mimicking scaffolds with heterogeneous composition and architecture. Constructs can be generated either from user-defined 3D computer-aided design (CAD) models or patient-derived imaging data such as CT or MRI scans. 3D bioprinting has achieved remarkable milestones in vascularizing engineered tissue constructs in the forms of microchannels[23], blood vessel-like threads[24], and tubular constructs[25]. Bioink-containing biomaterials and living cells is an essential component of bioprinting. It can be cross-linked or stabilized immediately after bioprinting to generate the final architecture of the intended construct[26]. Numerous extracellular matrix (ECM)-derived biomaterials such as fibrin, collagen, and Matrigel, are considered suitable bioinks to sustain the angiogenic activity of endothelial cells (EC) [27–29]. However, direct printing of these biomaterials remains challenging due to their unsuitable mechanical and thermo-responsive properties that cause poor shape fidelity and structural collapse. Recently, new breakthroughs have been accomplished to address this issue through the employment of a supporting medium. For instance, Noor et al. used alginate microparticles as supporting medium to print thick cardiac structures in high resolution[30]. Likewise, Lee et al. reported the utility of gelatin microparticles as a supporting bath for the printing of native, unmodified collagen with increased fidelity by an order of magnitude[31]. Cell-laden microbeads are inherently a cell-friendly particulate formulation, and thereby possess the potential to serve as a supporting medium for the printing of vasculature in a modular tissue engineering approach[32, 33]. More importantly, the microbeads can self-assemble into macroscopic tissues after printing instead of being degraded, leading to the creation of a complex TEBG containing vasculature with great fidelity independent of the stiffness of the bioink. In this study, we aim to explore the potential of combining modular tissue engineering with supporting medium-based printing to create a vascularized TEBG (Fig. 1). We prepared collagen microbeads loaded with human bone marrow derived mesenchymal stem cells (BMSCs) using a microfluidic approach. The cell-laden microbeads were characterized for dynamic size change, cell activity, osteogenesis, as well as self-assembly properties. We also tested the printability of a bioink loaded with endothelial cells derived from human induced pluripotent stem cells (iPSC-ECs) inside a microbead bath. The distribution and outgrowth of the iPSC-ECs were carefully assessed. Finally, we also explored and validated the feasibility of this system in the patterning of a centimeter-sized vertical pattern of iPSC-ECs.

Figure 1. Summary of the workflow.

In this study, microbeads loaded with human bone marrow derived mesenchymal stem cells (BMSCs) were prepared by a microfluidic approach and then characterized. The printability of endothelial cells derived from human induced pluripotent stem cells (iPSC-ECs) was tested inside a microbead bath as an attempt to create a centimeter-sized bone graft pre-vascularized with pattern of iPSC-ECs.

2. Material and Methods

2.1. Cell Culture and Differentiation

Frozen human bone marrow mesenchymal stem cells were purchased from RoosterBio (MSC-001), thawed in 37 °C water bath, and then cultured in standard growth medium (DMEM-high glucose, 10% fetal bovine serum (FBS), 100 units/ml penicillin and 100 mg/ml streptomycin, Gibco). At 80% confluence, cells were detached with 0.25% trypsin in 1 mM EDTA (Gibco) and passaged. Cells at passages 3–5 were harvested for all experiments. For osteogenesis, the cells were treated with differentiation medium consisting of DMEM, 10% FBS, 50 μg/ml ascorbate, 10 mM β-glycerophosphate, and 0.1 μM dexamethasone for 21 days.

Human Episomal induced pluripotent stem cells (Thermofisher A18945) were maintained in mTeSR plus medium (STEMCELL Technologies #05825) on cell culture plastic pre-coated with Geltrex (Thermofisher A1413201) with daily medium change, and passaged every 3–4 days. In order to acquire the endothelial phenotype, the iPSCs were firstly treated with DMEM/F12 medium supplemented with N2, B27, 5 μM CHIR99021 (Thermofisher), and 25 ng/ml BMP-4 (Peprotech 120–05) for 3 days for mesoderm induction, and then re-seeded with StemPro-34 medium (Thermofisher 10639011) containing 50 ng/ml VEGF-A (Peprotech 100–20), and 5 μM forskolin (LC laboratories F-9929) for 4 days. The differentiating iPSCs were then maturated in endothelial cell growth medium (EGM) supplemented with 100ng/ml VEGF-A. The efficiency of endothelial cell (EC) differentiation was assessed by flow cytometry to identify the CD31-positive population (BD FACSCelesta), along with PCR assays with TaqMan probes (Thermofisher) targeting key EC markers including CD31, CD34, CD144, von Willebrand factor (VWF), and eNOS. The acquisition of EC functionality was evaluated by acetylated low density lipoprotein (AcLDL) uptake assay and tube formation assay (Thermofisher). All characterizations were performed based on the manufacturer’s protocol.

2.2. Preparation of Cell-Laden Beads

Bone marrow MSCs were used at passage 3–5, re-suspended in neutralized, ice-cold collagen type I solution (Corning 354249), loaded into a 5 ml sterile syringe, and kept on ice until use. The cell suspension was pumped into the primary inlet of the microfluidic device at 50–80 μl/min, and sterile mineral oil (Sigma-Aldrich) was pumped into secondary inlets at 700–800 μl/min in the orthogonal direction to pinch the collagen solution flow and generate droplets at the junction point. The collagen droplets were cured in the channel with a 37 °C water bath and then collected at the outlet of the device (Fig. 2A). The collected beads were gently shaken and allowed to precipitate for 5 min to separate from the oil phase before washing 3 times in 1× PBS to remove the residual oil, and then maintained in standard MSC medium (RoosterBio KT-001) for up to 3 weeks.

Figure 2. Characterization of MSC-laden collagen microbeads.

A: Schematic of MSC-laden collagen bead preparation using a microfluidic system. B: The collagen beads prepared in this manner were homogenously spherical and evenly populated by MSCs. C: The ethidium homodimer-1-based tracking of cell death showed a consistently low and declining percentage of dead cells. D: Cell metabolic activity significantly increased 3 and 7 days after cell encapsulation (N=3). E, F: Bead diameter dropped significantly from 2.85 mm to 862 μm on average (N=3). G, H: The elongation of loaded MSCs was visualized by phalloidin staining and quantified over 14 days (N=3). *: p<0.05; **: p<0.01.

2.3. Cell Viability Analysis

Three separate assessments were performed to gain a comprehensive understanding of cell viability post-encapsulation. Firstly, the cell titer assay (96 Aqueous One Solution Cell Proliferation Assay, Promega G3582) was performed on days 0, 3, 7 and 14 to determine the metabolic activity of the cell-laden microbeads. Three beads were incubated in 250 μl of working solution for 3 hours in each independent trial and then the absorbance measured spectrophotometrically at 490 nm using a microplate reader (Tecan Spark). To visualize the distribution of cells in freshly prepared microbeads on day 0, the encapsulated MSCs were labelled with calcein acetoxymethyl (calcein-AM,1:200 in growth medium) for 30 minutes, and then visualized by fluorescence microscopy. To track cell death inside beads over the culture period of two weeks, cell-laden beads were stained with ethidium homodimer-1 (EthD-1, 1:500 diluted in medium) and then countered stained by 0.2% Hoechst 33342 Solution (Thermofisher 62249) on days 0, 3, 7 and 14. The samples were then washed 3 times in 1× PBS and viewed using fluorescence microscopy.

2.4. Fluorescent Staining

For F-actin staining, cell-laden beads were fixed in 10% formalin for 30 min, washed in PBS and then treated in 0.1% Triton X-100 for 30 min. The permeabilized samples were incubated with Alexa Fluor 488 phalloidin (Invitrogen A12379) and then counterstained with Hoechst before imaging by a fluorescent confocal microscope (Olympus FV3000 laser scanning confocal system). To visualize collagen turnover, 5-FAM conjugate collagen hybridizing peptide (F-CHP, 3Helix FLU60), was employed. Briefly, on days 1, 3, and 7, the beads were fixed with 4% paraformaldehyde for 4 hrs and washed in PBS. F-CHP working solution of 5 μM was prepared in PBS and then heated for 5 min at 80 °C in a water bath. The reagent, cooled to room temperature, was immediately added onto the fixed cell-matrix constructs to incubate overnight at 4 °C. Following staining, the beads were rinsed with PBS under gentle shaking for 30 min 3 times before counterstaining with Hoechst at room temperature. The staining was examined by the abovementioned fluorescent confocal microscope.

2.5. Tracking of Bead Aggregation

3 days after the initial bead preparation, 10 cell-laden beads were transferred into each well of a round bottom ultra-low attachment plate and maintained in growth medium supplemented with ascorbate-2-phosphate (50 mg/mL; Sigma-Aldrich) for 7 days to accelerate matrix deposition. The bead aggregate was harvested and subjected to 1 minute of physical agitation by vortex. Meanwhile, an equal number of aggregated beads were treated with trypsin-EDTA (Thermofisher) for 3 minutes before shaking. The number of beads remaining adherent to each other after shaking was counted and defined as the degree of aggregation. Moreover, 100–120 beads were transferred to each well of an un-treated 96 well plate along with a mold prepared by a stereolithography-based 3D printer (EnvisionTEC Perfactory IV). The mold consisted of a central post 1 mm in diameter and a flat base (Fig. 3). Beads were loaded around the post and cultured to induce self-assembly.

Figure 3. Investigation of the self-assembly property of cell-laden beads.

A: Strategy of bead aggregation. B, C: Outcome of bead aggregation with or without trypsin-EDTA treatment (N=3). The degree of aggregation was defined by the number of beads remaining adherent to each other after shaking. D, E: 3D printed molds were used to create a cylindrical construct with a central channel via bead aggregation. F: H & E staining confirmed the retention of the architecture (left). A higher magnification examination of the slides manifested the enrichment of cells and ECM deposition at the interface of adjacent beads (upper right), while spindle-shaped cells remained residing inside the beads (lower right). G,H: Porosity of the bead aggregates resulted in improved diffusion of FITC-labelled BSA compared to a casted collagen control. *: p<0.05; **: p<0.01.

2.6. Real-Time PCR Analysis

At each time point, total RNA of MSCs loaded in microbeads was extracted using the RNeasy extraction Kit (Qiagen) according to the manufacturer’s protocol. First-strand cDNA was synthesized with random primers using a cDNA synthesis kit (Invitrogen). Quantitative real-time PCR was performed using Taqman probes (Thermofisher) for osteogenesis markers runx2 (Hs01047973_m1) and osteocalcein (Hs01587814_g1) on a real-time PCR system (7900HT Applied Biosystem, Life Technology) and then analyzed by comparative Ct quantification (delta-delta Ct method). The expression level of each gene was normalized to GAPDH (Hs02786624_g1) as an endogenous control.

2.7. Histology

Cell laden beads were washed 3 times in PBS, fixed for 1 hour in 10% phosphate buffered formalin at room temperature, dehydrated in serial ethanol solutions followed by xylene treatment, and embedded in paraffin wax. The embedded samples were sectioned at 15 μm thickness using a microtome (Leica RM2255), deparaffinized in xylene, rehydrated, and stained with hematoxylin and eosin (H&E), or Alizarin red for visualization according to established protocols.

2.8. Patterning of iPSC-ECs in a Supporting Medium

iPSC-ECs were pre-labelled with an orange fluorescent signal by incubation with Celltracker™ Orange CMRA Dye (Thermofisher) in EGM for 30 min and then re-suspended in the bioink before printing. For the outgrowth study shown in Fig. 5, the labelled iPSC-ECs were suspended in 2 mg/ml fibrin on ice, and then loaded into the central channels of the constructs by pipetting. For the test of printing (Fig. 6), the bioink used in this study was composed of 4% gelatin-methacrylate (GelMA) and 5 mg/ml fibrin and kept at 4 °C at all times during the printing to prevent pre-gelation. The bioink with or without 1×10⁷/mL iPSC-ECs was extruded inside a bath of microbeads from a pre-cooled 22-gauge needle and cured by exposure to UV for 30 seconds. The distribution and outgrowth of fluorescently labelled iPSC-ECs was visualized by confocal microscopy (Olympus FV3000 laser scanning confocal system). Cell-free beads and bioink of the same material setting were used to optimize the extrusion parameters on a six-axis 3D bioprinter (BioAssemblyBot).

Figure 5. Loading of iPSC-ECs into the modular construct.

A: Human induced pluripotent stem cells (iPSCs) are a promising patient-specific cell source capable of endothelial commitment after two weeks of stepwise differentiation. B: The iPSCs adopted an endothelial phenotype, expressing the endothelial marker CD31, uptaking acetylated low density lipoprotein (acLDL) and forming tube-like structures on Matrigel. C: Constructs were formed with a central channel, either through casting or from microbeads. iPSC-ECs were loaded into the central channel, and the outgrowth of the cells was observed using confocal microscope. D: Top view of day 0 and day 3 samples. The iPSC-ECs were pre-labelled with CellTracker Orange and appear red in the images. The dotted white line denotes the border between the central channel (where ECs were loaded) and the surrounding collagen beads or casted collagen. On day 0, both casted and microbeads-based samples displayed a distinct iPSC-EC-rich center (left, insets are the gross appearance of the constructs from top view). On day 3, we noted the radial outgrowth of numerous iPSC-ECs from the border of the channel in both groups (white arrows). In contrast, network-like structures consisting of iPSC-ECs were found in areas distal to the center in the microbead group, but not in the cast group (right). E: Schematic depicting how the interstitial space between microbeads covered by MSCs may act as a “highway” that accelerated outgrowth compared to cast samples.

Figure 6. 3D patterning of endothelial cells.

A: Schematic of the printing strategy. B: Side-view images demonstrating that a support-free linear pattern of the hydrogel (red) can be deposited at any desired height inside the bead bath. The dotted lines indicate the border of the wells. C: When iPSC-ECs were loaded into the bioink and patterned in the bead bath (upper left, gross appearance), we noted a high percentage of live cells (upper right) and a distinct homogeneous, linear distribution of cells inside the print without any noticeable spreading into the surrounding space (bottom). D: Top-view confocal image of samples 3 days post-printing. Outgrowth of iPSC-ECs and formation of cord-like structures (indicated by white arrows) from the initial iPSC-EC locus was observed. The dotted white line denotes the border between the central channel and the surrounding beads. E: More complicated patterns can be created using the same extrusion-based printing.

2.9. Statistical Analysis

Data are presented as mean ± standard deviation. All quantitative assays were performed no fewer than thrice independently (N equals to the number of independent tests in figure legends). In each test, all samples were treated and analyzed separately in duplicate, and data were averaged. One-way ANOVA with Tukey’s post-hoc test and Student’s t-test was performed in Prism 8 (GraphPad) to determine statistical significance. Significance was considered at p < 0.05.

3. Results

3.1. Characterization of MSC-Laden Collagen Microbeads

MSC-laden collagen beads were prepared using a microfluidic system developed in our lab (Fig. 2A). The collagen beads prepared in this manner displayed homogenous size and spherical shape (Fig. 2B, left), as well as an even distribution of live MSCs within each bead (Fig. 2B, right). The ethidium homodimer-1-based tracking of cell death over a time span of 7 days indicated a consistently low and declining percentage of dead cells compared to the overall cell nuclei present in beads visualized by the counter staining of Hoechst (Fig. 2C). Moreover, the MTS-based cell titer assay revealed a significant increase in cell metabolic activity 3 and 7 days after cell encapsulation, respectively (Fig. 2D). These two findings jointly validated the cell compatibility of the initial bead preparation technique and the cell sustaining property of the beads. Contraction/deformation of collagen constructs by the seeded cells are commonly observed[34]. Since we aim to explore the potential of MSC-laden beads as a supporting medium for printing, it is pivotal to understand the shape change kinetics to ensure structural stability of the beads prior to printing. The macroscopic view of the beads (Fig. 2E) showed an evident transformation over 14 days from semi-transparent beads on the millimeter scale to opaque particles on the scale of several hundred microns. The structural integrity and spherical shape were well preserved. Two weeks of culture led to a significant drop in bead diameter from 2.85 mm to 862 μm on average (Fig. 2F), and plateaued for up to another two weeks without noticeable collapse or degradation (data not shown). In the meantime, through phalloidin staining that visualized the cell cytoskeleton, we noted the elongation of MSCs to adopt spindle shaped morphology over 7 days (Fig. 2G). We quantified the cell morphological change via analyzing the cell aspect ratio, defined as the long axis divided by short axis of each cell, and observed that the kinetics reflect the trend of elongation (Fig. 2H). This finding suggests the mechano-sensing activity of MSCs within the beads applied static tension onto the surrounding ECM, contributing to the contraction of the collagen beads.

3.2. Investigation of the Self-Assembly Property of Cell-Laden Beads

Sufficient integration of micro-tissues is the key to successful modular tissue engineering. In this study, we investigated the ability of cell-laden beads to self-assemble into macroscopic tissue. When cell-laden beads were cultured in round bottom ultra-low attachment wells for 7 days (Fig. 3A), robust aggregates were formed that could withstand physical agitation by vortex (Fig 3B, middle). In contrast, cell-free beads showed no sign of aggregation (Fig. 3B, left). Interestingly, we noted the dissociation of cell aggregates after trypsin-EDTA treatment (Fig. 3B, right). The degree of aggregation, defined by the number of beads remaining adherent to each other after shaking, indicated a significant decline of integration in the trypsin-EDTA treated group (Fig. 3C). We thereby concluded that the integration is mediated by intrinsic cell–cell and cell–ECM interactions at the interface between adjacent beads. The self-assembly properties of the MSC-laden beads enabled us create centimeter-sized constructs in a bottom-up manner. A pre-fabricated mold with a vertical post was placed in the well prior to the placement of beads (Fig. 3D and 3E inset) to create a cylindrical construct with a central channel via bead aggregation over 7 days of culture (Fig. 3E). Histological analysis afterward in the transverse view confirmed the retention of the conduit in the center (Fig. 3F, left). An examination of the slides at higher magnification indicated the enriched cells and ECM deposition at the interface of adjacent beads as well as a portion of spindle-shaped cells remaining inside the beads (Fig. 3F, right). Despite the macroscopic structural integrity of the construct, histological analysis revealed plenty of interstitial micro-space in every plane inside the construct that is absent in collagen constructs prepared by casting (Fig. 3G). This finding implies the potential difference in permeability between these two construct settings when perfused by the flow of media. To test this hypothesis, we perfused the central channel of constructs with FITC-labelled BSA, and tracked the diffusion of BSA to the surrounding tissue in real time. The fluorescent image series suggest remarkable differences in the pattern of diffusion between casted collagen and bead aggregates (Fig. 3H). The intensity of fluorescence across the Y axis in the images were plotted to quantitatively reflect the degree of diffusion from the center of the channel. The flatter bell-shaped curve in the bead aggregate group suggests a markedly higher degree of diffusion at all the time points analyzed (Fig. 3H).

3.3. Assessment of Cell Activity in Beads

It is widely reported that MSCs are capable of remodeling the microenvironment to favor subsequent cell behavior. In this study, we assessed the cell-matrix interaction. Collagen hybridizing peptide (CHP) is a unique peptide that specifically binds collagen helices through the repeating Gly-x-y sequence. CHP has a strong specificity to hybridize with denaturing or synthesizing collagen chains to form the triple helical structure. It has negligible affinity to intact collagen molecules due to lack of binding sites (Fig. 4A). The outcome of the hybridization on our cell-laden beads revealed active matrix remodeling by MSCs that may involve both degradation of the existing collagen presented by beads and the deposition of neo-synthesized collagen from cells (Fig. 4B). We also confirmed cell responsiveness to exogenous osteogenic cues: When the beads were treated with standard osteogenic medium for three weeks, PCR analysis suggested significantly upregulated expression of runx2 and osteocalcein at day 7 and day 14, respectively (Fig. 4C). Alizarin red staining on the sectioned microbeads indicated evident calcium deposition at day 21 compared to the control beads maintained in growth medium (Fig. 4D).

Figure 4. Assessment of cell activity in beads.

A: Collagen hybridizing peptide (CHP) has a strong specificity to hybridize with denaturing or synthesizing collagen chains to form the triple helical structure. B: Active matrix remodeling by MSCs in beads revealed by CHP staining. C: When the beads were treated with standard osteogenic medium, PCR analysis suggested significantly upregulated expression of runx2 and osteocalcein at day 7 and day 14, respectively (N=3). D: Alizarin red staining on the sectioned microbeads indicated evident calcium deposition at day 21 compared to the control beads maintained in growth medium. *: p<0.05.

3.4. Patterning iPSC-ECs inside the Modular Construct

We evaluated the interaction between ECs and aggregated MSC-microbeads as an attempt to initiate vascular structure formation within the engineered graft. Human induced pluripotent stem cells (iPSCs) are a promising patient-specific cell source capable of endothelial commitment to vascularize bone grafts[35]. In this study, after two weeks of stepwise differentiation, more than 90% of the iPSCs adopted an endothelial phenotype, expressing the endothelial marker CD31, forming tube-like structures on Matrigel and uptaking acetylated low density lipoprotein (acLDL) (Fig. 5A, B). The iPSC-ECs were pre-labelled with Celltracker Orange, suspended in fibrin, and loaded into the central channels of the constructs (Fig. 5C). The top view of the day 0 samples displayed a distinct iPSC-EC-rich center (Fig. 5D) as opposed to the surrounding body of the construct. Three days later, we noted the radial outgrowth of numerous iPSC-ECs from the border of the channel in both groups (Fig. 5D, white arrows). In addition to single cell level outgrowth, network-like structures consisting of a large number of iPSC-ECs were found in areas more distal to the center in the microbead group (Fig. 5D, right column). We speculate that the interstitial space between microbeads covered by MSCs paved a “highway” that led to this accelerated outgrowth (Fig. 5E). Taken together, our data suggests the feasibility of creating a pattern of active iPSC-ECs inside the modular engineered construct.

3.5. Patterning Vasculature via 3D Printing

In addition to all abovementioned cell-sustaining and self-assembly properties, the MSC-laden microbeads are also an inherently hydrophilic microparticulate formulation that possesses the potential to support free-form 3D printing of angiogenic bioink. In this study, we tested the potential of printing a bioink consisting of fibrin, methacrylated gelatin (GelMA) and iPSC-ECs inside a bath of MSC-laden microbeads. We firstly confirmed the feasibility of this printing strategy using a simplified set up: a cell free, ice-cold mixture of fibrin and GelMA was manually extruded atop a layer of microbeads, cured by UV, and then covered by another layer of microbeads (Fig. 6A). In this way, we were able to create a support-free linear pattern of the hydrogel at any desired height inside the bead bath by simply adjusting the level of the bath underneath the printout (Fig. 6B). When iPSC-ECs pre-labelled with CellTracker Orange were loaded into the bioink and extruded in the same experimental set up, we noted a high percentage of live cells and a distinct homogeneous, linear distribution of cells inside the printout without any noticeable free spreading into the surrounding space (Fig. 6C). Moreover, outgrowth of iPSC-ECs and subsequent formation of cord-like structures from the initial iPSC-EC locus was observed after three days of culture in the standard EGM medium (Fig. 6D), indicating the potential to vascularize the entire construct. Using the same extrusion-based printing, we were also able to customize more complicated patterns than basic geometrical shape (Fig. 6E).

4. Discussion

In this study, we explored a strategy to create a vascularized TEBG via combining modular tissue engineering and supporting medium-based printing (Fig. 1). In particular, we prepared collagen microbeads loaded with BMSCs using a microfluidic approach. The size change, cell activity, osteogenesis, and the self-assembly property of these beads were finely characterized. The promising findings from these studies led to the test of printability of iPSC-ECs inside a microbead bath, which revealed high survival and rapid outgrowth of iPSC-ECs. Finally, we explored and validated the feasibility of this system for the 3D patterning of a centimeter-sized vertical pattern of iPSC-ECs.

Although the initial size of the beads collected from the device was two to three millimeters, cell-mediated contraction markedly reduced the bead size as a function of cell seeding density (Supplementary Fig. 1). The impact of static tension on encapsulated cells is therefore worth studying in the future[34]. Moreover, since the particle size of bath medium is a determinant factor for the porosity of macrotissue and the precision of the subsequent bath print, cell seeding density should be carefully planned in accordance with the desired permeability and resolution of the print.

Human iPSCs are a promising autologous cell source for regenerative medicine. Compared to embryonic stem cells, iPSCs can be derived from patient somatic cells and thereby circumvent concerns of immune rejection and the ethical concern related to tissue harvest from embryos. In this study, we used an established protocol with slight modifications to accomplish the highly efficient endothelial commitment of iPSCs in order to vascularize our bone graft (Supplementary Fig. 2)[36]. Our lab is also exploring the potential of iPSCs to differentiate into MSCs, which can then give rise to functional osteoblasts for bone regeneration (Supplementary Fig. 3)[37–39]. Using MSCs and ECs both derived from patient specific iPSCs for this modular tissue engineering approach will be our next goal to pursue, which can minimize the risk of immunogenicity when implanted in patients.

Robust assembly of microtissues is the prerequisite to any bottom-up modular tissue engineering. To date, various integration technologies have been reported, including chemical cross-linkers[40] and UV-activated gluing agents[41, 42]. In our study, the cell-laden beads aggregated into macrotissues in the absence of any exogenous chemical treatment, but via intrinsic cell–cell and cell–ECM interactions taking place at the interface between adjacent beads. Our findings are consistent with the previously reported high rate of cell vitality and robust microtissue integration using a similar assembly strategy on a different platform of microparticles[43]. There are several limitations of our method to be addressed in future studies. The essential time of cell culture (up to seven days) for the aggregate to form inevitably prolonged the time span of graft preparation. The development of an interpenetrating hydrogel network consisting of collagen and another biocompatible polymer that enables instantaneous bead assembly via secondary crosslinking after collagen curing will be a worthy research goal to accomplish for future studies. Moreover, the mechanical strength of the current construct is not high enough for clinical implantation. This shortcoming may be overcome by confining the construct in a secondary, 3D printed scaffold that possesses the desired mechanical properties and microstructure of bone tissue, such as the 3D printed bone-like scaffolds composed of polycaprolactone (PCL) previously created in our lab[44].

Unlike other reports of supporting medium-based 3D printing, where the microparticles were removed to acquire the printed construct, we instead preserved the cell-laden beads to constitute the parenchyma of the graft that presented microporosity for augmented permeability (Fig. 3). The printing enabled the creation of customizable 3D patterns of iPSC-ECs with good precision and fidelity. To our best knowledge, this fabrication strategy has not been reported before. Sufficient vascularization of the engineered construct is the key to successful TEBGs. In this study, the outgrowth from the iPSC-ECs into the surrounding space indicates the survival and activity of iPSC-ECs patterned by our method. However, the lack of assessment of tubule/vessel formation remains a limitation of the current study. Small animal cranial bone defect models will be employed in our future study to fully assess the bone formation, functional vascularity, and anastomosis to host tissue. The construct size in our current experiment setting is still limited by the nutrient supply to the center of the construct before sufficient angiogenesis of patterned iPSC-ECs. This limitation can be addressed by integrating a vascular conduit in the center of the construct to constantly perfuse the entire tissue using a bioreactor, as we have previously reported[14].

In conclusion, our study suggests the feasibility of 3D printing vasculature inside a bath of cell-laden microbeads to create centimeter-sized modular tissue engineered constructs. This study holds great potential to advance the development of a much-needed artificial bone graft with autologous patients’ cells, customizable architecture, and pre-existing vasculature.

Statement of Significance.

Current clinical strategies to repair severe bone damage, such as autografts and allografts, are hindered by many limitations. The ability to engineer large-volume and functional bone grafts is therefore critical for clinical bone repair. Modular tissue engineering is a promising bio-fabrication strategy to create engineered bone grafts in a bottom-up manner. In this study, the cell-laden microbeads self-assembled into macro-tissues, enabling precise control over the cell spatial distribution in macroscale and offering great flexibility for the volume and architecture of grafts. Moreover, the cell-laden microbeads served as a supporting medium for the patterning of iPSC-derived endothelial cells to vascularize the graft. Our approach opens new possibilities to create complex customizable tissue engineered bone grafts containing printed vasculature with great fidelity independent of the stiffness of the bio-ink.