Microfluidic technology may also be utilized to optimize culture conditions for difficult-to-grow stem cells (e.g., References 19,20,21). At present, most research to identify niche factors is based on constant, static exposure of stem cells to extrinsic stimuli. Since stem cells can be regulated in a time-dependent manner [e.g., the circadian rhythm (22)], microfluidic platforms may be powerful tools to elucidate such time-dependent processes. Microfluidic systems, in contrast to static culture systems, allow rapid medium exchange and culture condition switching at desired time points during an experiment. An example of this was demonstrated by King and colleagues, who developed a microfluidic chip for the high-throughput variation of temporal stimuli (23). An implemented ‘flow-encoded switch’ enabled the simultaneous delivery of many different temporal profiles of a cellular stimulus. By controlling the pressure difference between the buffer and protein flow, variable pulse train widths, lengths, and frequencies were achieved nearly independently. With this device, the effect of tumor necrosis factor–alpha (TNFα)–induced apoptosis of hepatoma cells was investigated. By varying the recovery time after a heat shock treatment in a single experiment, the authors demonstrated that recovery was maximal after 4–7 h, supporting the dual role of TNFα in promoting both cell survival and apoptosis.
Cellular fate changes due to high shear stresses are potential confounding factors in microfluidic cultures. Conversely, perfusion is crucial for the delivery of fresh nutrients and growth factors. Consequently, microfluidic chips have been designed to optimize perfusion rates in stem cell cultures. Such chips typically consist of a network of parallel channels having different hydrodynamic resistances to vary the perfusion rate in each microchannel (24,25). Interestingly, such chip experiments have indicated that ESC culture can be improved at higher perfusion rates (21,25), one study showing a linear correlation of proliferation with perfusion rate (25). ESC growth reached a maximum at high flow rates, exchanging the chip medium in <5 min, and seemed not to depend on shear stress but rather limited by medium or cytokine supply.
Despite the apparent advantages of microfluidic technology and its successful application in many areas— including system biology (26), crystallography (27), and bioanalytics (28,29)—the impact of microfluidic technology in stem cell research thus far has been moderate. This may be explained by some unmet challenges regarding its use in a standard laboratory. One of the major limitations of microfluidic platforms is the so-called ‘world-to-chip’ problem: how to connect the picoliter microfluidic scale to our microliter world. Efforts to facilitate this down-scaling (or up-scaling) of over six orders of magnitude has involved automated on-chip valves and pumps that could be actuated using pressure and external solenoid valves (20,21) or via a Braille display incorporated into a microfluidic chip (30). Although these external valves can be controlled digitally and multiplexing allows the fabrication of arrays of thousands of individually addressable chambers with a limited number of on-chip valves (31), the reliable delivery of hundreds or thousands of different microenvironments to microfluidic devices remains a bottleneck (32). Parallelization of inputs can significantly increase throughput, as Maerkl and Quake demonstrated by combining spotting technology and microfluidics (26), and can possibly alleviate current limitations of microfluidic technology for stem cell applications.Mimicking the spatially controlled display of niche signals
The stem cell niche is not a homogeneous microenvironment as emulated by microarrays spotted on rigid cell culture substrates, but rather a spatially well-defined 3-D heterogeneous structure (Figure 1D). Thus, the impact of the biophysical niche properties on stem cell fate (including its mechanical properties and its 3-D architecture) should not be underappreciated (2,3,4). Cellular- or subcellular-scale approaches to simplify this spatially complex system and to study structural aspects of the niche can be implemented using microfabrication or microfluidics. For example, McBeath et al. observed that the differentiation of human mesenchymal stem cells (hMSCs) cultured on plastic depended on cell shape (33). By confining single hMSCs to small or large fibronectin islands under mixed osteogenic and adipogenic culture conditions, they proved that restriction of cell spreading area induced adipogenesis whereas cells that could spread widely preferentially differentiated into osteoblasts (Figure 3A). This fate switch was mediated by RhoA, a small GTPase involved in regulation of the actin cytoskeleton, and its downstream effector ROCK. The role of RhoA in lineage commitment was so central that cells transfected with constitutively active or dominant negative RhoA differentiated into osteoblasts or adipocytes, respectively, independent of soluble factors.