Microfluidics Enabled Bottom-Up Engineering of 3D Vascularized Tumor for Drug Discovery

Abstract

Development of high-fidelity three-dimensional (3D) models to recapitulate the tumor microenvironment is essential for studying tumor biology and discovering anticancer drugs. Here we report a method to engineer the 3D microenvironment of human tumors, by encapsulating cancer cells in the core of microcapsules with a hydrogel shell for miniaturized 3D culture to obtain avascular microtumors first. The microtumors are then used as the building blocks for assembling with endothelial cells and other stromal cells to create macroscale 3D vascularized tumor. Cells in the engineered 3D microenvironment can yield significantly larger tumors in vivo than 2D-cultured cancer cells. Furthermore, the 3D vascularized tumors are 4.7 and 139.5 times more resistant to doxorubicin hydrochloride (a commonly used chemotherapy drug) than avascular microtumors and 2D-cultured cancer cells, respectively. Moreover, this high drug resistance of the 3D vascularized tumors can be overcome by using nanoparticle-mediated drug delivery. The high-fidelity 3D tumor model may be valuable for studying the effect of microenvironment on tumor progression, invasion, and metastasis and for developing effective therapeutic strategy to fight against cancer.

Keywords: angiogenesis; core−shell microcapsule; tumor microenvironment; vascularization; vasculogenesis.

Figure 1

Characterization of proliferation and gene expression of avascular 3D μtumors. (a) A differential interference contrast (DIC) image of a typical microcapsule showing its core-shell morphology and a scanning electron microscopy (SEM) image showing collagen (1.5 mg/ml) fibers in the microcapsule core. Alg: alginate and Col: collagen. (b) Distribution of the total and core size of the microcapsules together with the distribution of the number of cells in them. (c) Typical phase contrast and fluorescence (live/dead) images of the encapsulated cells in microcapsules with a 1.5 mg/ml collagen core, showing the cell proliferation over 10 days. (d) Elastic modulus of different core ECMs. (e) Quantification of relative cell proliferation represented by the size of aggregates on day 10 obtained from culturing the cells in different collagen core ECMs in the microcapsules. The size is normalized to that of the 3 mg/ml collagen core condition. (f) Quantitative RT-PCR data showing the effect of matrix stiffness on the expression of mesenchymal (VIMENTIN and CXCR4) and epithelial (E-CADHERIN) phenotype markers of cells in the μtumors. The symbol * denotes p < 0.05. Scale bar: (a) 100 μm for the DIC image or 5 μm for the SEM image and (c) 100 μm.

Figure 2

Assembly of μtumors in the microfluidic perfusion device to form vascularized 3D tumor. (a) Live/dead staining of the vascularized tumor on different days showing high cell viability. HUVECs without green fluorescence protein (GFP) were used. Arrow and arrowhead represent microcapsule shell and μtumor, respectively. (b) Phase contrast and fluorescence images (4×) showing vessel formation on day 4 with (1) μtumors encapsulated in microcapsules, (2) empty microcapsules, and (3) μtumors without microcapsules (the alginate hydrogel shell was dissolved by perfusing the samples with 75 mM sodium citrate solution for 5 min after one day of culturing the sample). HUVECs with GFP were used. (c) Time-lapse micrographs at 10× showing the progression of vessel formation within 4 days. Extensive vascularization was observed on days 3 and 4 when the μtumors were present, whereas vascularization was minimal with empty microcapsules. Moreover, removing the alginate hydrogel shell after one-day culture further facilitates vascularization. Arrow and arrowhead represent microcapsule shell and μtumor, respectively. (d) The vessel structure visualized by staining for ACTIN filament, CD31, and cell nuclei. Cross-sectional images (i, ii, and iii) demonstrate the presence of lumen in the vessels. HUVECs without GFP were used. Scale bar: (a and c) 100 μm, (b) 200 μm, (d) 50 μm and for cross-sectional images: 20 μm.

Figure 3

In vivo tumorigenicity of the 3D-engineered system of encapsulated μtumors, HUVECs, and hADSCs in collagen. (a) Photographs of mice and tumors obtained on 14 days after subcutaneously injecting single 2D-cultured MCF-7 cancer cells aone (condition 1), mixture of single hADSCs, HUVEC, 2D-cultured MCF-7 cancer cells (condition 2), the 3D-engineered system (condition 3), and the 3D-engineered system with dissolution of alginate at 1 day after injection (condition 4) in collagen into athymic nude mice. The total number of cells were the same for all the conditions. (b) Weight of the tumors collected on day 14 post subcutaneous injection. (c) Representative histology (H&E) images of tumors, showing blood vessels with red blood cells (RBCs, arrow heads). (d) Quantitative measurement of the density of blood vessels in tumors obtained from the four conditions. (e) Confocal images of immunofluorescent staining for human CD31 (hCD31, green) and cell nuclei (blue) in the tumors. HUVECs without GFP were used. The symbol * denotes p < 0.05. Scale bar: (a) 5 mm, (c) 40 μm, (e) 100 μm.

Figure 4

Drug response of the 3D vascularized tumor. (a) Viability of 2D-cultured cancer cells, 3D avascular μtumors (3D-A), and 3D vascularized tumors (3D-V) after 4 days of incubation or treatment with free doxorubicin hydrochloride (DOX). Also shown is the IC50 of free DOX for the three different tumor models. The symbol * denotes p < 0.05 when compared to 2D-cultured cancer cells, and # denotes p < 0.05 when compared with 2D-cultured cancer cells and 3D avascular μtumors. (b) A schematic illustration of the LC60S-DI nanoparticles consisting of lipid (L), fullerence (C60), silica (S), DOX (D) and indocyanine green (ICG or I). (c) Size distribution and (d) surafce zeta potential of the nanoparticles determined by dynamic light scattering (DLS). (e) Viability of all cells in 3D-V after 4 days of treatment with LC60S-DI with various effective DOX concentrations. (f) IC50 of free and nanoencapsulated DOX for the destruction of 3D vascularized tumors. The symbol $ denotes p < 0.05 when comparing between the free and nanoparticle-encapsulated DOX.

Scheme 1

Schematic illustration of the bottom-up approach for creating 3D vascularized human tumor. (a) A non-planar microfluidic encapsulation device is used for encapsulating cancer cells in core-shell microcapsules and the cells are cultured in the microcapsules for 10 days to form micro-tumors (μtumors, less than ~200 μm in radius). Mineral oil infused with calcium chloride, aqueous sodium alginate solution (to form the microcapsule shell), aqueous collagen solution (with or without cells) to form the microcapsule core, and aqueous extraction solution are pumped into the device via inlets I1, I2, I3, and I4, respectively. The aqueous phase (containing core-shell microcapsules) and oil exit the device from outlets O1 and O2, respectively. (b) A microfluidic perfusion device is used to assemble the μtumors and stromal cells including endothelial cells for perfusion culture to form 3D vascularized tumor. The μtumors in core-shell microcapsules are assembled together with human umbilical vein endothelial cells (HUVECs) and human adipose-derived stem cells (hADSCs) in collagen hydrogel in the microfluidic perfusion device. The alginate shell of the microcapsules is dissolved to allow cell-cell interactions and the formation of 3D vascularized tumor in the microfluidic perfusion device under perfusion driven by hydrostatic pressure. Unit for the dimensions of micro-pillars and sample chamber: mm; P: pressure; ρ: density; g: gravitational acceleration; and h: height of medium column linked to the reserviors.