Micro-dissected tumor tissues (MDTs) are maintained alive on chip for several days and show promising results for personalized medicine applications.

We present a microfluidic platform for automatic multi-size spheroid formation within constant volume hanging droplets (HDs) from a single inlet loading of a constant cell concentration.

Microphysiological systems hold the promise to increase the predictive and translational power of in vitro substance testing owing to their faithful recapitulation of human physiology. However, the implementation of academic developments in industrial settings remains challenging. We present an injection-molded microfluidic microtissue (MT) culture chip that features two channels with 10 MT compartments each and that was designed in compliance with microtiter plate standard formats. Polystyrene as a chip material enables reliable, large-scale production and precise control over experimental conditions due to low adsorption or absorption of small, hydrophobic molecules at or into the plastic material in comparison with predecessor chips made of polydimethylsiloxane. The chip is operated by tilting, which actuates gravity-driven flow between reservoirs at both ends of every channel, so that the system does not require external tubing or pumps. The flow rate can be modulated by adjusting the tilting angle on demand. The top-open design of the MT compartment enables efficient MT loading using standard or advanced pipetting equipment, ensures oxygen availability in the chip, and allows for high-resolution imaging. Every channel can be loaded with up to 10 identical or different MTs, as demonstrated by culturing liver and tumor MTs in the same medium channel on the chip.

This work focuses on modelling design and operation of “microfluidic sample traps” (MSTs). MSTs regroup a widely used class of microdevices that incorporate wells, recesses or chambers adjacent to a channel to individually trap, culture and/or release submicroliter 3D tissue samples ranging from simple cell aggregates and spheroids, toex vivotissue samples and other submillimetre-scale tissue models. Numerous MST designs employing various trapping mechanisms have been proposed in the literature, spurring the development of 3D tissue models for drug discovery and personalized medicine. Yet, there lacks a general framework to optimize trapping stability, trapping time, shear stress, and sample metabolism. Herein, the effects of hydrodynamics and diffusion-reaction on tissue viability and device operation are investigated using analytical and finite element methods with systematic parametric sweeps over independent design variables chosen to correspond to the four design degrees of freedom. Combining different results, we show that, for a spherical tissue of diameterd < 500 μm, the simplest, closest to optimal trap shape is a cube of dimensionsw equal to twice the tissue diameter:w = 2d. Furthermore, to sustain tissues without perfusion, available medium volume per trap needs to be 100× the tissue volume to ensure optimal metabolism for at least 24 hours.

As 3D in vitro tissue models become more pervasive, their built-in nutrient, metabolite, compound, and waste gradients increase biological relevance at the cost of analysis simplicity. Investigating these gradients and the resulting metabolic heterogeneity requires invasive and time-consuming methods. An alternative is using electrochemical biosensors and measuring concentrations around the tissue model to obtain size-dependent metabolism data. With our hanging-drop-integrated enzymatic glucose biosensors, we conducted current measurements within hanging-drop compartments hosting spheroids formed from the human colorectal carcinoma cell line HCT116. We developed a physics-based mathematical model of analyte consumption and transport, considering (1) diffusion and enzymatic conversion of glucose to form hydrogen peroxide (H₂O₂) by the glucose-oxidase-based hydrogel functionalization of our biosensors at the microscale; (2) H₂O₂ oxidation at the electrode surface, leading to amperometric H₂O₂ readout; (3) glucose diffusion and glucose consumption by cancer cells in a spherical tissue model at the microscale; (4) glucose and H₂O₂ transport in our hanging-drop compartments at the macroscale; and (5) solvent evaporation, leading to glucose and H₂O₂ upconcentration. Our model relates the measured currents to the glucose concentrations generating the currents. The low limit of detection of our biosensors (0.4 ± 0.1 μM), combined with our current-fitting method, enabled us to reveal glucose dynamics within our system. By measuring glucose dynamics in hanging-drop compartments populated by cancer spheroids of various sizes, we could infer glucose distributions within the spheroid, which will help translate in vitro 3D tissue model results to in vivo.