3D microfabricated bioreactor with capillaries
- First Online:
- Cite this article as:
- Xia, C. & Fang, N.X. Biomed Microdevices (2009) 11: 1309. doi:10.1007/s10544-009-9350-4
- 334 Views
We present in this paper the implementation of an innovative three dimensional (3D) microfabrication technology coupled with numerical simulation to enhance the mass transport in 3D cell culture. The core of this microfabrication technology is a high-resolution projection micro stereolithography (PμSL) using a spatial light modulator as a dynamic mask which enables a parallel fabrication of highly complex 3D microstructures. In this work, a set of poly (ethylene glycol) microfabricated bioreactors are demonstrated with PμSL technology. We observed both experimentally and numerically the regulation of metabolism and the growth of yeast cells by controlling the density of micro-capillaries. Further development of these 3D microfabricated bioreactors is expected to provide artificially constructed tissues for clinical applications.
Reconstructive surgeries are performed to recover the function and appearance of damaged tissues, especially following major cancer resections and traumas. It is estimated that more than one million reconstructive surgery procedures are performed by plastic surgeons every year. The development of reconstructive surgery (Dunn and Watson 2001) has proved the success of free flaps as reconstructive tissues for recipients. A free flap is a block of tissue with inherent microcirculatory network, usually is transferred from a patient’s own body close to the defective site (Taylor and Palmer 1987). However, the nature of sacrificing one part of a body for another limits the application of free flaps in practice. Therefore alternative tissue sources for reconstructive surgery are desired. The field of tissue engineering (Berthiaume and Yarmush 2002) introduces the exciting possibility of replacing damaged body parts with new ones customized to the specific needs of the recipient, independent of the availability of donor sources. A number of tissue-engineered products are currently used clinically, such as Integra™, Carticel™, and Apligraf™. Further development of cell biology, micro-technology, and biomaterial science will provide new opportunities to create larger and more complex artificial tissues.
One of the major obstacles towards the creation of large complex 3D artificial tissues is the lack of microcirculatory system at the early stage of tissue culture (Ruben et al. 2005). The time scale for neo-vascularization is in the order of days (even with growth factors) and the time scale for cell death from hypoxia is in the order of hours. Therefore without capillary perfusion, the metabolism during cell growth cycles will eventually exhaust the supply of nutrient and oxygen from the external environment and the embedded cells suffer from the lack of nourishment, creating a bottleneck for the growth of thick (>1 mm scale) 3D tissues. Studies (Sutherland et al. 1986; Martin et al. 1999) confirmed that cells in a tissue were poorly cultured when they were further than ~400 μm away from external nutrient sources. As a matter of fact, in real tissues, most of cells stay within a distance of about 100 μm from nearby capillaries (Berthiaume and Yarmush 2002). Several research groups have developed methods to enhance the mass transport in tissue culture by taking advantage of current microfabrication technologies. For example, by inserting and extracting nylon strands and tubing, straight artificial blood vessels were created to allow the continuous perfusion of culture medium (Neumann et al. 2003). However, the assembly of many discrete micro blood vessels into an inner-connected 3D network for nutrient perfusion will not be practical. Griffith, et al (Griffith et al. 1997) created 3D channels by 3D printing technology. Unfortunately, the resolution of this technology was only 200 μm, which is much larger than a capillary dimension (<20 μm). Silicon microfabrication technologies and molding were also able to create two dimensional micro channels for enhanced mass transport (Borenstein et al. 2002). A recent study (Levenberg et al. 2005) shown a multi-culture system consisting of myoblasts, embryonic fibroblasts and endothelial cells co-seeded on highly porous, biodegradable polymer scaffolds could induce the endothelial vessel networks. However the scaffolds used in this work were less than 1 mm thick, which is still within the diffusion distance of external nutrients. Nevertheless, the three dimensional nutrient transport in thick (>1 mm) tissue culture still remains a hurdle in tissue engineering.
To enhance the transport and exchange of nutrients and wastes for constructing thick artificial tissues, a novel three dimensional microfabrication technology, projection micro stereolithography (PμSL) (Sun et al. 2005), is introduced for the design and the fabrication of vascularized micro bioreactors. We show in this work such micro fabricated bioreactors, coupled with mass transport simulation, can dramatically enhance the nutrition and growth of cultured cells through capillary networks. This microfabrication method brings several unique advantages to the field of tissue engineering: first, the capability of PμSL to build truly 3D sophisticated microstructures with very fine spatial resolution at micron scale; second, a significantly shortened design cycle enabled by high fabrication speed (500 layers in a couple of hours); finally, the choice of biocompatible and biodegradable polymers offers flexibility on fabricating implantable pre-vascularized scaffolds for different tissue cultures (Ratner and Bryant 2004; Hou et al. 2004).
2 Microfabrication and materials
The basic capabilities of our projection micro stereolithography system
Max. sample size
2 µm (in plane or horizontal) 1 µm (off plane or vertical)
4 mm3/hour (limited by viscosity of resins)
15 mm × 11 mm ×10 mm
3 Vascularized micro bioreactor
Where Vmaxis the maximal uptake rate and KM is the metabolite concentration when the uptake rate is half of the maximum. In Michaelis-Menten kinetics, the consumption behavior follows first order kinetics at low concentration. That means the consumption rate is proportional to the concentration. As the concentration of metabolites increases, the consumption behavior will become zero order kinetics gradually. At a certain point, the cell is saturated and the intake of metabolites reaches a plateau.
Our simulation indicates that the bottleneck of effective glucose transport is the permeability of polymer capillaries. The glucose concentration drops off more than 95% after diffusing through the capillary wall. The simulation shows that if the center to center distance of the capillaries is set to 120 µm and the wall of the capillary is 10 µm, then the inner radius of the capillary has to be larger than 20 µm to ensure that all the yeast cells in the bioreactor has a high enough (>830 nmol/mL) glucose concentration to stay in the mixed repiro-fermentative metabolism and produce ethanol (Fig. 5). This configuration corresponds to 80.2 capillaries/mm2 if capillaries are in hexagonal arrangement. By increasing the inner radius of the capillary, not only the perfusion of the culture medium is increased, but also the gap between capillaries is decreased. It is equivalent to increase the density of the capillary. When the inner radius is 20 µm, the lowest glucose concentration in the bioreactor is 880 nmol/mL. Further decreasing the inner radius of capillaries will decrease the glucose concentration in the bioreactor and force some yeast cells start to consume ethanol (Otterstedt et al. 2004; Verduyn et al. 1984). The biomass growth becomes much slower than at a higher glucose concentration. Two experiments at different points of the simulation curve are also shown in Fig. 5. Experiment A is in the Phase I region that the glucose concentration in the bioreactor is much higher than 830 nmol/mL. Experiment B is at the cutoff region between Phase I and Phase II. We observed dramatic difference of the biomass production (Fig. 5(b) and (c)). Actually in experiment A (Fig. 5(b)), the yeast cells filled the whole bioreactor and even pushed the cells on top out of the bioreactor during culture. When the bioreactor was removed from the culture chamber, the top layers of yeasts were washed away. However, experiment B (Fig. 5c) shows the exact amount of yeast that we achieved. According to the simulation, in experiment B, the yeast cells should also have fully filled the bioreactor as shown in experiment A. We contribute this error to a low glucose concentration in the bioreactor right after the cell seeding. In experiment B, the capillaries were sparser than that in experiment A. Without cells blocking the way, the glucose diffusing from the capillary quickly escaped from the bioreactor, causing a too low glucose concentration to keep the proliferation of yeast cells. A more detailed model to better capture the real time growth of yeast cells in bioreactors will be our future efforts.
4.1 Measurement of effective diffusion coefficient of ethanol in PEG membranes
We measured the effective diffusion coefficient of ethanol in PEG membranes using the Kaufmann-Leonard method (Kaufmann and Leonard 1968). Two identical stirred compartments were filled with deionized water and 50 w% ethanol water solution. These two compartments were separated by a 200-μm PEG membrane which was fabricated by UV-polymerizing a layer of PEG diacrylate solution between two glass slices. The motor stirred at a rate of 1 Hz. After 10 min, the concentration of ethanol in deionized water was measured using QED alcohol test kit (OraSure Technologies, Inc.). Since the change of ethanol concentration in both compartments was very small compared to 50 w% ethanol solution, we assumed the diffusion mass flux = -De×C0 /d was constant during the measurement. Here De is the effective diffusion coefficient; C0 is the concentration of ethanol in 50 w% ethanol solution and d is the thickness of PEG membranes. Therefore De can be calculated from equation δm = De × C0 × δt/d. Here δm is the amount of ethanol in deionized water during a time period of δt.
4.2 Yeast cell culture
The yeast we used was diploid strain INVSc1 (Invitrogen). Before yeast culture, the bioreactors were fabricated using PμSL and kept in 100% ethanol for 24 h and biology grade water for 24 h to remove the residue monomer and initiator, also to increase the permeability of the capillaries. The yeast suspension in 1.5 mL microcentrifuge tube was moved from -70°C freezer and left in 20°C room temperature for 20 min before they were seeded in the micro bioreactor using 0.1–10 µL micro pipette. The number of seeded yeast was around 80. The micro bioreactor was placed in the reaction chamber (1 in. × 0.5 in. × 0.5 in.) filled with DPBS. Two steel micro tubes with OD 400 μm penetrated the chamber side walls and were connected to the micro bioreactor inside as shown in Fig. 4(c). The chamber was covered with quarter inch thick transparent PLEX sheet to prevent possible contamination. The yeast culture medium YPD (1 g yeast extract (Difco), 2 g Peptone (Difco), 2 g D-glucose, 100 mL distilled water) was delivered at a flow rate of 0.5 mm/s through the capillaries in the micro bioreactor. The culture chamber is kept in a humidified incubator at 30°C for 45 h. The DPBS solution in the chamber was replaced with fresh one every 6~10 h to remove the ethanol in the chamber. The glucose concentration in the replaced DPBS was measured using GlucCell™ glucose monitoring system. Finally, the incubated micro bioreactor was removed from the chamber and dried in air at room temperature for 1 h before spattering coating and SEM observation.
Projection Micro-Stereolithography (PμSL) promises rapid design and manufacturing of advanced micro bioreactors by offering a unique opportunity to culture tissues in vitro. By integrating high density micro capillary channels within the micro bioreactors, the mass transport can be enhanced by advection to withstand the increasing demand of oxygen and nutrients during cell growth. Simulation based on glucose diffusion model showed that the bottleneck of effective transport was the diffusivity of the polymer material of the capillary. The glucose concentration dramatically decreased after diffusing through the wall of the capillary. The S. cerevisiae yeast cell culture well verified the simulation prediction. Not only is this model applicable for glucose, but also for the transport of other metabolites for different cells. Our simulation modeling can predict how far the nutrients transport into cell layers. With the predicted transport distance, we can precisely control the density of the polymer capillary to ensure that all the cells in the micro bioreactor are in healthy nutrient state.
This project was supported by the Gauthier Exploratory Research Foundation.