Micro-bioreactor arrays for controlling cellular environments: Design principles for human embryonic stem cell applications Elisa Cimetta a, Elisa Figallo a, Christopher Cannizzaro b, Nicola Elvassore a,* Gordana Vunjak-Novakovic c,* a Department of Chemical Engineering, University of Padova, Padova, Italy b Department of Biomedical Engineering, Tufts University, Medford, MA, USA c Department of Biomedical Engineering, Columbia University, 622 West 168th Street, VC12-234, New York, NY 10032, USA a r t i c l e i n f o Article history: Accepted 16 October 2008 Available online 24 October 2008 Keywords: Bioreactor Human embryonic stem cells Hydrogel Microfluidics Steady-state Dynamic conditions Time scale Transport phenomena a b s t r a c t We discuss the utilization of micro-bioreactor arrays for controlling cellular environments in studies of factors that regulate the differentiation of human embryonic stem cells. To this end, we have designed a simple and practical system that couples a microfluidic platform with an array of micro-bioreactors, and has the size of a microscope slide [E. Figallo, C. Cannizzaro, S. Gerecht, J.A. Burdick, R. Langer, N. Elvassore, G. Vunjak-Novakovic, Lab Chip 7 (2007) 710���719]. The system allows quantitative studies of cells cultured in monolayers or encapsulated in three-dimensional hydrogels. We review the operating requirements for studies of human embryonic stem cells (hESCs) under steady-state and dynamic conditions, and the related control of the mass transport and hydrodynamic shear. We describe the design and fabrication of the individual bioreactor components, and the crite- ria for selecting the bioreactor configuration and operating parameters, based on the analysis of the characteristic times and scales of reaction, convection and diffusion. To illustrate the utility of the bio- reactor, we present a ������case study��� of hESC cultivation with detailed experimental methods and rep- resentative biological readouts. �� 2008 Elsevier Inc. All rights reserved. 1. Introduction Novel cell culture technologies developed in recent years mimic the in vivo cellular microenvironments with an increasing fidelity [2], through improved control and the provision of cascades of multiple regulatory factors. Miniaturization of the culture systems is an important step towards accurate control of the cultured cells and tissues. Some of the most interesting outcomes come from the optimization and accurate use of microfluidic platforms [3,4]. Small transport distances are key for enabling fast-responses to environmental stimuli in studies involving spatial and temporal gradients of factors. Since its first appearance decades ago, micro- fluidics have been adapted to many different applications several excellent reviews give insights into the operating principles for system configurations of interest [5���8]. Microscale technologies were designed for applications ranging from studies at a single cell level [9] to the recreation of more complex 3D tissues [10] and the development of diagnostics platforms [11,12]. High-tech platforms involving integrated microdevices such as micro-valves, injectors, pumps or mixers [13] are also being considered for use in live cell experimentation. Our laboratory has been interested in developing a micro-biore- actor array that would allow high throughput studies of the most challenging cell source: human embryonic stem cells (hESCs). To this end, we have designed an efficient, multiplexed device that couples a microfluidic platform with an array of micro-bioreactors and allows quantitative studies of hESCs in two-dimensional (2D) and three-dimensional (3D) settings [1]. This system enables quan- titative measurement of performance and accurate control over the culture microenvironment with a relative simplicity, not requiring additional integrated devices such as micro-valves, pumps or mix- ers. In this paper, we first summarize the design requirements for systems of this kind. We review the operating requirements for studies of cultured cells under steady-state and dynamic condi- tions, and the related control of the mass transport and hydrody- namic shear. The design and fabrication of the individual bioreactor components and the system assembly are described in detail. Then we describe the design specifications for important operating parameters and the principles for their optimization at the microscale level that are based on the analysis of transport rates and characteristic times of the involved phenomena. Finally, we present a ������case study��� of hESC cultivation, with detailed exper- imental methods and representative biological readouts. 1046-2023/$ - see front matter �� 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ymeth.2008.10.015 * Corresponding authors. Fax: +1 212 305 4692 (G. Vunjak-Novakovic). E-mail addresses: nicola.elvassore@unipd.it (N. Elvassore), gv2131@columbia. edu (G. Vunjak-Novakovic). Methods 47 (2009) 81���89 Contents lists available at ScienceDirect Methods journal homepage: www.elsevier.com/locate/ymeth

2. Overall requirements 2.1. Biomimetics Our body is composed of highly organized tissues formed by multiple cell types and sustained through complex structure���func- tion relationships at many hierarchical levels. Short and long range communication between single cells and/or tissues occurs through an enormous cascade of signaling pathways, most of which involve mechanisms, intermediate steps and connections that are not fully understood. Ideally, biological studies should be performed under conditions that are controllable and at the same time capture the complexity of the native environment. 2.2. Scale When it comes to scaling down to the biologically relevant lengths and time scales, the conventional methodologies of the so-called ������flat biology��� fail. Standard culture methods involving cell culture in Petri dishes cannot be representative of the real state of physiologic systems and therefore often result in unrealistic and uncontrollable biological readouts. The relatively large volume of medium in contact with cells and the batch-wise operation associ- ated with the periodic medium exchange does not allow for the generation and control of precise patterns of stimulation such as spatio-temporal gradients. Petri dishes are therefore not suitable for use in studies of either steady-state or dynamic system The batch processes have unpredictable time scales, are diffusion-lim- ited and, most important, are intrinsically uncontrollable. Macro- scale observations of averaged cell/tissue properties cannot be reliable in defining complex biological systems and must thus be replaced by more precise and realistic microscale observations. In this respect, the microscale culture technologies help overcome most of the above limitations, and to operate at the characteristic time and length scales of biological phenomena. 2.3. Time constants At the microscale���inside the microchannels and culture cham- bers���the dominating forces change and differ from the most known, macroscale ones. Flow is always laminar (Reynolds num- bers are well below the turbulence threshold, and in most cases Re 100) the inertial forces are therefore dominated by viscous forces, and the transport is dominated by molecular diffusion or by convective regime of well-defined hydrodynamic profile. Reducing the characteristic dimensions to a microscale level allows a more rigorous control of the operating parameters involved, due to the very short transport distances, which are in turn associated with very short time constants. As a result, biological responses are not limited anymore by the slow kinetics of physical phenomena. 2.4. Multi-parametric analysis The miniaturization of the system also results in the reduction of the amounts of cells, culture media and supplements and there- by helps reduce the cost and time involved in cell culture, and en- ables the high throughput quality of the data. The possibility to control multiple factors, molecular and physical, allows multi- parametric analyses, thus again reducing both time and cost of experimentation. 2.5. Imaging compatibility Most micro-bioreactors are optically transparent and fully com- patible with conventional imaging techniques. They allow online analyses and real-time experimentation such as time-lapse moni- toring the state of the system and the time-course of specific phenomena. 3. Specific requirements 3.1. Steady-state conditions Perfused systems offers the advantage of working under well- defined and stable steady-state conditions. Standard culture sys- tems, such as Petri dishes, can be compared to batch reactors in which conditions are precisely defined only at time zero, and then continuously vary until next medium exchange. Working at stea- dy-state conditions maintains, by definition, the parameters of interest at their constant levels. Precise perturbations of the sys- tem may then be introduced to investigate the dynamic biological responses of the cells. 3.2. Fast transients in space and time It is often necessary to accurately modulate transients in space and time to recreate precise stimulation patterns in the form of concentration gradients, and deliver particular signals. Working with small volumes (order of a few lL) generally leads to fast-re- sponse systems. This means that the switch between defined sets of conditions can be fast and still controllable. For example, oper- ating parameters can be changed by a simple variation of the med- ium flow rate. At the microscale, one can design the system to provide tightly controlled, orderly conditions of flow and mass transport. 3.3. Control of mass transport and flow shear Due to the well-defined geometry, short transport distances, and fast transients, transport phenomena occurring in microde- vices can be more easily subjected to theoretical analysis and pre- cise control than those in larger scale systems. Accurate predictions of the velocity gradients and shear stresses are indeed very important because of the profound effects of flow environ- ment on biological systems. Again, the versatility of microscaled systems, allows decoupling the effects of mass transport phenom- ena such as the generation of specific concentration patterns from physical phenomena such as the application of mechanical forces. 4. Technical description of the micro-bioreactor platform This section contains the description of the main components of the micro-bioreactor array culture system: the cell culture cham- ber, the microfluidic platform, the gas exchange unit, the tubing and connections, and the flow circulation system. The described elements can be visualized in Fig. 1, keeping in mind that their application can be extended to different configurations. Briefly, pa- nel A shows a top view of the micro-bioreactor array assembly with its functional units, which will be treated more in detail in the following paragraphs. Panel B presents an image of the culture system set up in the incubator, while the lateral views in panel C clarify the layered structure of the micro-bioreactor array exempli- fying the two configurations used. 4.1. Culture chamber���2D and 3D settings The cell culture chambers were designed to provide (i) culture of a statistically significant number of cells for biological assays, (ii) easy and rapid access to the chamber for application of stan- dard protocols (i.e., coating, fixing and staining), (iii) the mainte- 82 E. Cimetta et al. /Methods 47 (2009) 81���89