We developed a heterogeneous breast cancer model consisting of immortalized tumor and fibroblast cells embedded in a bioprintable alginate/gelatin bioink. The model recapitulates the in vivo tumor microenvironment and facilitates the formation of multicellular tumor spheroids, yielding insight into the mechanisms driving tumorigenesis.
The cellular, biochemical, and biophysical heterogeneity of the native tumor microenvironment is not recapitulated by growing immortalized cancer cell lines using conventional two-dimensional (2D) cell culture. These challenges can be overcome by using bioprinting techniques to build heterogeneous three-dimensional (3D) tumor models whereby different types of cells are embedded. Alginate and gelatin are two of the most common biomaterials employed in bioprinting due to their biocompatibility, biomimicry, and mechanical properties. By combining the two polymers, we achieved a bioprintable composite hydrogel with similarities to the microscopic architecture of a native tumor stroma. We studied the printability of the composite hydrogel via rheology and obtained the optimal printing window. Breast cancer cells and fibroblasts were embedded in the hydrogels and printed to form a 3D model mimicking the in vivo microenvironment. The bioprinted heterogeneous model achieves a high viability for long-term cell culture (> 30 days) and promotes the self-assembly of breast cancer cells into multicellular tumor spheroids (MCTS). We observed the migration and interaction of the cancer-associated fibroblast cells (CAFs) with the MCTS in this model. By using bioprinted cell culture platforms as co-culture systems, it offers a unique tool to study the dependence of tumorigenesis on the stroma composition. This technique features a high-throughput, low cost, and high reproducibility, and it can also provide an alternative model to conventional cell monolayer cultures and animal tumor models to study cancer biology.
Although 2D cell culture is widely used in cancer research, limitations exist as the cells are grown in a monolayer format with a uniform concentration of nutrients and oxygen. These cultures lack important cell-cell and cell-matrix interactions present in the native tumor microenvironment (TME). Consequently, these models poorly recapitulate physiological conditions, resulting in aberrant cell behaviors, including unnatural morphologies, irregular receptor organization, membrane polarization, and abnormal gene expression, among other conditions1,2,3,4. On the other hand, 3D cell culture, where cells are expanded in a volumetric space as aggregates, spheroids, or organoids, offers an alternative technique to create more accurate in vitro environments to study fundamental cell biology and physiology. 3D cell culture models can also encourage cell-ECM interactions that are critical physiological characteristics of the native TME in vitro1,4,5. The emerging 3D bioprinting technology provides possibilities to build models that mimic the heterogeneous TME.
3D bioprinting is derived from rapid prototyping and enables the fabrication of 3D microstructures that are capable of mimicking some of the complexities of living tissue samples6,7. The current bioprinting methods include inkjet, extrusion, and laser-assisted printing8. Among them, the extrusion method allows the heterogeneity to be controlled within the printed matrices by precisely positioning distinct types of materials at different initial locations. Therefore, it is the best approach to fabricate heterogeneous in vitro models involving multiple types of cells or matrices. Extrusion bioprinting has been successfully used to build auricular shaped scaffolds9, vascular structures10,11,12, and skin tissues13, resulting in high printing fidelity and cell viability. The technology also features versatile material selections, the ability to deposit materials with cells embedded with a known density, and high reproducibility14,15,16,17. Natural and synthetic hydrogels are frequently used as bioinks for 3D bioprinting due to their biocompatibility, bioactivity, and their hydrophilic networks that can be engineered to structurally resemble the ECM7,18,19,20,21,22,23.Hydrogels are also advantageous since they can include adhesive sites for cells, structural elements, permeability for nutrients and gases, and the appropriate mechanical properties to encourage cell development24. For instance, collagen hydrogels offer integrin anchorage sites that cells can use to attach to the matrix. Gelatin, denatured collagen, retains similar cell adhesion sites. In contrast, alginate is bioinert but provides mechanical integrity by forming crosslinks with divalent ions25,26,27,28.
In this work, we developed a composite hydrogel as a bioink, comprised of alginate and gelatin, with similarities to the microscopic architecture of a native tumor stroma. Breast cancer cells and fibroblasts were embedded in the hydrogels and printed via an extrusion-based bioprinter to create a 3D model that mimics the in vivo microenvironment. The engineered 3D environment allows cancer cells to form multicellular tumor spheroids (MCTS) with a high viability for long periods of cell culture (> 30 days). This protocol demonstrates the methodologies of synthesizing composite hydrogels, characterizing the materials' microstructure and printability, bioprinting cellular heterogeneous models, and observing the formation of MCTS. These methodologies can be applied to other bioinks in extrusion bioprinting as well as to different designs of heterogeneous tissue models with potential applications in drug screening, cell migration assays, and studies that focus on fundamental cell physiological functions.
1. Preparation of the Materials, Hydrogel, and Cell Culture Materials
2. Measurements of Rheological Properties of Hydrogels
3. Scaffold Design, Cell-laden Hydrogel, and 3D Printing Models
4. Viability and Spheroid Formation Experiments on the Hydrogel Disks.
5. Scanning Electron Microscopy (SEM)
The temperature sweep shows a distinct difference of the A3G7 precursor at 25 °C and 37 °C. The precursor is liquid at 37 °C and has a complex viscosity of 1938.1 ± 84.0 mPa x s, which is validated by a greater G" over G'. As the temperature decreases, the precursor undergoes physical gelation due to the spontaneous physical entanglement of the gelatin molecules into a tri-helix formation29,30. Both the G' and the G" increase and converge at 30.6 °C, indicating a sol-gel transition. The moduli continue to increase, with decreased temperature, reaching 468.5 ± 34.2 Pa of G' and 140.7 ± 9.3 Pa of G" at 25 °C (Figure 1a). Based on the results of the temperature sweep, we chose 25 °C as the printing temperature. The gelation kinetics experiment simulates the temperature change that occurs during the sample preparation and handling (i.e., removing the sample from the 37 °C water bath and placing it into the 25 °C bioprinter chamber). G', G", and |η*| increase with time at the lowered temperature, and the sol-gel transition happens at ~ 17 min at 25 °C, when G' ≈ G" ≈ 64.3 Pa and the loss factor equals 1 (Figure 1b and 1c). The precursor continues to stiffen and reaches a printing window after 50 – 90 min of gelation. Within this printing window, the precursor can be smoothly extruded using a G25 cylindrical nozzle with a pressure of 200 kPa. The existence of yield stress implies a solid-like behavior of the material and raises the structural integrity after the extrusion to withstand its own weight31. The stress ramp recognizes the yield stress at different times of the gelation. Results show that the yield stress increases with the increasing gelation time. At 50 min of gelation, the yield stress reaches 325.9 Pa, assuring that the model is stable after the extrusion.
The printed propeller model is shown in Figure 2a. Confocal microscopy confirms the initial extrusion locations of both the MDA-MB-231-GFP and the IMR-90-mCherry cells (Figure 2b). The MDA-MB-231-GFP cells begin to develop spheroids 15 days into the culture period, followed by increased size and numbers of spheroids until day 30 (Figure 2c – 2f). Some of the IMR-90-mCherry cells also form agglomerations (Figure 2g – 2j). Noticeably, after 30 days of culture, IMR-90-mCherry cells are observed in the region initially occupied by the MDA-MB-231-GFP cells (Figure 2f), implying possible migration events in the model. Likewise, migrated MDA-MB-231-GFP cells can also be observed in the IMR-90-mCherry dominated region on day 30 (Figure 2j).
The disk model is shown in Figure 3a. Confocal microscopy confirms the homogeneous cell distribution within the disk (Figure 3b). The MDA-MB-231-GFP cells behave similarly in the disk model as they did when printed as a propeller model. Representative images are displayed in Figure 3c – 3f. SEM imaging reveals an MDA-MB-231-GFP spheroid-laden hydrogel after 21 days of culture (Figure 4a). The generation, and presence, of the spheroid is validated by comparing the SEM images with cell-free hydrogels (Figure 4b).
The cell viability of both cell types is demonstrated in Figure 4c. MDA-MB-231-GFP cells express a higher cell viability compared to IMR-90-mCherry cells. Interestingly, the MDA-MB-231-GFP cells exhibit an increased trend of viability before day 15, with a tripling of the number of viable cells compared to day 0. This is followed by a decrease on day 21, before recovering again on day 30. In contrast, the IMR-90-mCherry cells have minimal fluctuations in viability during the entirety of the culture period. The decreasing values in viability of the MDA-MB-231-GFP cells at day 21 correspond to a large MTCSs formation, which has developed necrotic cores.
Figure 1: Rheological characterization of the hydrogel precursor. (a) A temperature sweep shows the sol-gel transition at 30.6 °C. (b-c) The gelation kinetics of the A3G7 precursor shows an increase of G', G", and |η*| at time of gelation, and the sol-gel transition happens at approximately 17 min at 25 °C. (d) This panel shows the yield stress of the precursor versus the time of gelation. An increase of yield stress is observed with a longer gelling time. The results are shown in mean ± SD, n≥ 3. Please click here to view a larger version of this figure.
Figure 2: MCTS formation within a 3D bioprinted in vitro model consisting of IMR-90-mCherry myofibroblast and MDA-MB-231-GFP breast cancer cells. (a) These photographs show the bioprinted in vitro sample (left) and the CAD model (right). (b) This representative confocal time-lapse image shows the MDA-MB-231-GFP (green) and IMR-90-mCherry (red) cells bioprinted within the model. (c – f) These zoom-ins show the MDA-MB-231-GFP cell regions (white dotted boxes). (g – j) These zoom-ins show the IMR-90-mCherry cell regions (yellow dotted boxes). The scale bars are 2 mm in panel 2a, 1 mm in panel 2b, and 500 µm in panels 2c – 2j for selected areas, and the magnification is 10X. Capital "D" in the images means "days of culture". Please click here to view a larger version of this figure.
Figure 3: MCTS formation within a 3D MDA-MB-231-GFP-laden hydrogel disk. (a) This photograph shows a hydrogel disk. (b) This representative confocal image shows MDA-MB-231-GFP cells embedded in the composite. (c-f) These zoom-ins show the MDA-MB-231-GFP cell region (white dotted box in panel 3b) at (c) 0, (d) 7, (e) 15, and (f) 21 days of culture. The scale bars are 2 mm in panel 3a, 1 mm in panel 3b, and 500 µm in panels 3c – 3f for selected areas, and the magnification is 10X. Capital "D" in the images means "days of culture". Please click here to view a larger version of this figure.
Figure 4: Morphology and cell viability of MTCSs. (a) This SEM image shows an MDA-MB-231-GFP MCTS within the gel after 21 days of culture, showing small MCTS (arrow-pointed). (b) This SEM image shows an alginate/gelatin hydrogel without cells. The magnification is 350X. (c) This panel shows the cell viability of MDA-MB-231-GFP and IMR-90-mCherry during 30 days of culture within the hydrogel. The results were analyzed as the mean ± SD and statistically analyzed using a two-way ANOVA and a Bonferroni's post-test. p (*) < 0.05, n ≥ 3. The data was normalized to the cell density used to create the samples on day 0. Please click here to view a larger version of this figure.
Cell-laden structures can be compromised if contamination (biological or chemical) occurs at any point in the process. Usually, biological contamination is seen after two or three days of culture as a color change in the culture media or the bioprinted structure. Therefore, the sterilization (physical and chemical disinfection) is a key step for all the cell-related processes. Noteworthy, autoclaving gelatin changes its gelling properties, which made it gel slower in the trials we conducted. Therefore, we sterilized the alginate and gelatin power via UV exposure. Due to the very limited penetration capability of UV light, a very thin layer (< 0.5 mm) should be used. The concentration of alginate and gelatin can be arbitrarily changed to tune the mechanical and biological properties32. In the present work, we chose A3G7, because it provides desirable printability during its printing window, and the crosslinked hydrogel provides a high stability through the 30-day experiment. Heating to 60 °C while dissolving powder into DPBS helps to enhance the liquefication of gelatin, which eases the stirring. A lower temperature can also be used; however, a longer stirring time will then be required.
The rheological temperature sweep gives a sol-gel point at 30.6 °C, which is compatible with other publications33. Theoretically, any temperature higher than 30.6 °C can liquefy the precursor. We chose 37 °C to simplify the work, as cell culture medium is pre-warmed in a 37 °C water bath. Based on the gelation kinetics (before the precursor gels), there is approximately 17 min for cell mixing; after this time, mixing the cells and the hydrogel precursor is difficult due to the increased viscosity and moduli, causing a non-homogeneous cell-laden bioink. Therefore, we suggest handling the cell-mixing work within the first 10 min of gelling. The yield stress is not considered to be a material attribute that affects printability until recently34; regardless, it has been extensively recognized by polymer scientists as a marker of pasty/granular fluid31,35,36,37. The existence of yield stress helps to build up a structural integrity to withstand its own weight. A high yield stress can also require an increased startup pressure in extrusion; thus, this may result in damage to the cells due to excessive shear stress32. The model fidelity can be compromised if the extrusion process occurs outside of the optimal printing window. Common indicators of this problem are shown in the final bioprinted structure: it will either be too rough or too liquid at the edges. The A3G7 precursor exhibits an optimal printing window during 50 – 90 min of gelling that satisfies both the structural stability and the cell survival and can be used as the medium to build heterogeneous tumor models14.
The total height of the propeller model is 600 µm as it is generated from four layered 150 µm filaments. This design allows nutrient and gas exchange as cells are at most 300 µm from the surface contacting the media during the incubation. Higher models are achievable in fabrication but are bottlenecked by the limited diffusion distance of the nutrients in the hydrogel38, which can result in a low cell survival in the core region.
Different strategies have been developed to form MCTS in vitro, such as hanging drop, microfluidic chip, assembly, and bioprinting39. However, some of these methodologies can alter the cell physiology and biochemistry, generating different cell behaviors that occur in normal tumor tissue. For instance, the hanging drop method forces single cells to stay as cell aggregates through physical confinement40. Also, the chemical induction to form MCTS by a peptide addition can alter the biochemistry of the spheroids41. The hydrogel composite described here allows an MCTS formation by creating a biomimetic environment without external stressors. Starting out as single well-dispersed cells, after 7 days of cultures, cells reorganize as small MCTS (more than 6 cells per spheroid), increasing their size and quantity during 30 days of culture.
In the results of this research, we found that MDA-MB-231-GFP spheroids decrease the cell viability at day 21, correlating this phenomenon with the formation of large spheroids. A solid tumor is comprised of three different layers, whereby the external layer presents a high proliferation rate compared to the middle layer and the internal necrotic or senescent core of the spheroid18. This could explain the viability decrease in the results.
This limitations of this protocol are (1) bioink dynamics and (2) cell type compatibility. Bioink dynamics refers to the varying material properties during the gelation process, resulting in an optimal printing window beyond which the printed structure is defective. Cell type compatibility refers to the incapacity of certain cell types (cell line or primary culture) to exhibit a native in vivo behavior.
The A3G7 hydrogel achieves a high stability and cell viability. 3D bioprinting can be used to build 3D heterogeneous disease models with high-throughput, low cost, and high reproducibility as a more realistic alternative to traditional cell culture and small animal tumor models.
The authors have nothing to disclose.
Tao Jiang thanks the China Scholarship Council (201403170354) and McGill Engineering Doctoral Award (90025) for their scholarship funding. Jose G. Munguia-Lopez thanks CONACYT (250279, 290936 and 291168) and FRQNT (258421) for their scholarship funding. Salvador Flores-Torres thanks CONACYT for their scholarship funding (751540). Joseph M. Kinsella thanks the National Science and Engineering Research Council, the Canadian Foundation for Innovation, the Townshend-Lamarre Family Foundation, and McGill University for their funding. We would like to thank Allen Ehrlicher for allowing us to use his rheometer, Dan Nicolau for allowing us to use his confocal microscope, and Morag Park for granting us access to fluorescently labeled cell lines.
Sodium alginate | FMC BioPolymer | CAS-No: 9005-38-3 | Protanal LF 10/60 FT |
Gelatin | Sigma-Aldrich | G9391 | Type B gelatin from bovine skin |
Dubelcco's phosphate buffered saline (DPBS 1X) | Gibco | LS14190136 | 1×, w/o calcium, w/o magnesium |
Magnetic hotplate | Corning | N/A | Stirrer/hot plate model PC-420 |
50 mL centrifuge tubes | Corning | 352098 | Falcon® 50mL High Clarity PP Centrifuge Tube, Conical Bottom, Sterile |
Centrifuge | GMI | N/A | Sorvall RT6000D, GMI, USA |
Calcium chloride anhydrous | Sigma-Aldrich | C1016 | |
MilliQ water | Millipore | N/A | |
Millipore 0.22 µm filters | Millipore | SLGS033SB | Millex-GS Syringe Filter Unit, 0.22 µm, mixed cellulose esters, 33 mm, ethylene oxide sterilized |
Oscillation rheometer MCR 302 | Anton Paar | N/A | |
Rheometer measuring tool CP25 | Anton Paar | 79038 | Conical plate geometry for rheometer |
RheoCompass | Anton Paar | N/A | Software controlling rheometer MCR 302 |
Scanning electron microscope | Hitachi | N/A | SEM, Hitachi SU-3500 Variable Pressure |
Paraformaldehyde, 96%, extra pure | Acros Organics | 416785000 | |
Dulbecco modified eagle medium (DMEM) | Gibco | 11965092 | |
Antibiotic/Antimycotic solution (100X) stabilized | Sigma | A5955 | |
Fetal bovine serum | Wisent Bioproducts | 080-150 | |
Cell culture T-75 flasks | Sigma-Aldrich | CLS430641 | 75 cm2 TC-Treated surface treatment |
3D bioprinter BioScaffolder 3.1 | GeSiM | N/A | |
GeSim software | GeSiM | N/A | Software controlling BioScaffolder 3.1 |
10cc cartridge UV resist | EFD Nordson | 7012126 | |
End cap | EFD Nordson | 7014472 | |
Tip cap | EFD Nordson | 7014469 | |
Piston | EFD Nordson | 7012182 | |
Stainless nozzle G25 | EFD Nordson | 7018345 | |
Water bath | VWR | N/A | |
Agarose | Sigma-Aldrich | A9539 | Bioreagent, for molecular biology |
Costar 6-well plates | Corning | 3516 | TC-Treated Multiple Well Plates, Individually Wrapped, Sterile |
Confocal spinning disk inverted microscope | Olympus Life Science | N/A | Olympus IX83 |
MTS assay kit | Promega | G3582 | CellTiter 96® AQueous One Solution Cell Proliferation Assay |
Live/Dead viability cytotoxicity kit | Molecular Probes,ThermoFisher Scientific | L3224 | |
Trypsin 0.25/EDTA 1X | Gibco | 25200-072 | |
Corning 96-well plate | Corning | 3595 | Clear Flat Bottom Polystyrene TC-Treated Microplate, Individually Wrapped, with Low Evaporation Lid, Sterile |
Autoclave Tuttnauer | Heidolph Brinkmann | N/A | Heidolph Tuttnauer 2540E Autoclave Sterilizer Electronic Model with 4 Stainless Steel Trays, 23L Capacity |
Trypan blue | Invitrogen | T10282 | 0.4% solution |
Ethanol | Commercial Alcohols | P016EA95 | Greenfield Speciality Alcohols |
CO2 Incubator | Panasonic | N/A | MCO 19AIC-PA |
Lyophilizer | SP Scientific | N/A | Virtis Sentry 2.0 |
SolidWorks | Dassault Systems | N/A | A CAD software used to build demostrative propeller-like model |
MATLAB | The MathWorks | N/A | A programming software used to generate G-code for BioScaffolder 3.1 |