Pressure-Driven Microfluidic Perfusion Culture Device for Integrated Dose-Response Assays

Abstract

Cell-based assays are widely used in the various stages of drug discovery. Advances in microfluidic systems over the past two decades have enabled them to become a powerful tool for cell-based assays to achieve both reliability and high throughput. The interface between the micro-world and macro-world is important in industrial assay processes. Therefore, microfluidic cell-based assays using pressure-driven liquid handling are an ideal platform for integrated assays. The aim of this article is to review recent advancements in microfluidic cell-based assays focusing on a pressure-driven perfusion culture device. Here, we review the development of microfluidic cell-based assay devices and discuss the techniques involved in designing a microfluidic network, device fabrication, liquid and cell manipulation, and detection schemes for pressure-driven perfusion culture devices. Finally, we describe recent progress in semiautomatic and reliable pressure-driven microfluidic cell-based assays.

Introduction

The cost of drug development has increased exponentially in accordance with Eroom’s law, ¹ and the attrition rate in clinical trials is increasing year by year. ² In addition, bridging the gap between experiments in animal models and clinical trials is difficult because of the inherent differences between animals and humans, making extrapolation of results difficult. Therefore, in vitro assays using cultured cells are becoming an increasingly important tool. High-throughput screening systems using a microplate and dispensing robot are widely used for cellular assays in drug development research. ³ Recent advances have been made in the application of inkjet and acoustic systems for liquid handling, and assay throughput is increasing. In addition, technological advances now allow more information to be obtained from a single assay. ⁴ Further advances in microfluidic technology are expected to realize even more efficient throughput.

The drawbacks for using traditional cell-based assays for screening in drug development are the differences between the environment surrounding the cells and the animal body and the fact that the cultured cells have lost many of the functions found in the body. ⁵ These limitations have the potential to be overcome by the application of microfluidic technology to drug discovery applications by miniaturized assays and to increase experimental throughput and reliability in drug discovery applications.^6–8

In the past decade, microfluidic devices have been progressively used as versatile tools for many types of cell-based assays.^9–13 One of the key advantages of using microfluidics for cell-based assays is the miniaturization of each assay. Miniaturized high-throughput and high-content cell-based assays have been reported^14–17 and applied to analyses of liver drug metabolism and toxicity.^18,19 Assay miniaturization provides the opportunity for creating a variety of microenvironments, such as soluble factors and extracellular matrixes, within the microfluidic devices.^20,21 The fabrication of microfluidic devices with biomimetic tissues is another approach for cell-based assays.^22–24 This approach has been reported for lung,^25,26 renal, ²⁷ cardiac, ²⁸ artery, ²⁹ gut, ³⁰ and corneal tissue ³¹ and is expected to increase the reliability of in vitro assays.

Microfluidic devices have the advantage of reducing the amount of valuable sample and reagents required for cell-based assays and can be fabricated with engineered tissues to increase the reliability of results. The application of microfluidic devices to high-throughput industrial drug discovery has led to the development of the pressure-driven perfusion culture microchamber array chip. ³² Prior to the development of pressure-driven perfusion culture devices, microfluidic chips required the connection of tubes and external pumps (e.g., syringe pump and peristaltic pump) so that culture media could flow into the microfluidic network. This type of system proved to be too complicated and impractical for microfluidic chips to be used for cell-based assays designed to simultaneously screen multiple drug candidates. These issues were addressed by the pressure-driven perfusion culture microchamber array chip. ³²

Table 1 summarizes the comparison of technical features in the reported liquid delivery system for perfusion culture microfluidic devices. Most of the liquid delivery systems, including syringe pumping, pneumatic pressure, gravity force, and so on, require external tube connection. One exception that does not need tube connection is the surface tension–driven liquid delivery system, in which loaded liquid is spontaneously delivered by the intrinsic energy. Compared with these liquid delivery systems, the pressure-driven liquid delivery system requires only one to a few tube connections to pressure sources. In addition, fluid flow rate is easily calculated from the microchannel geometry and applied pressure. Considering these advantages of pressure-driven perfusion culture, we recently developed a user-friendly pressure-driven perfusion culture microchamber array chip for high-throughput drug dose-response assays. The design of this system allowed multiple liquids to be introduced under pressure into the microchip. We further improved this system by integrating a serial dilution microfluidic network to generate logarithmic concentration profiles. ⁵¹ We then demonstrated the efficacy of this microchamber array system by successfully evaluating the dose response of a model anticancer drug, paclitaxel, in HeLa cells. ⁴⁰ Here, we introduce recent technology developments of our pressure-driven perfusion culture microchamber array chip and highlight its simple interface and well-designed microfluidic network. This system has potential as an efficient platform for high-throughput cell-based microchip assays in drug discovery applications.

Table 1.

Comparison of Liquid Delivery Systems for Perfusion Culture Microfluidic Devices.

Liquid Delivery System	Complexity of Microfluidic Devices	External Equipment	Requirement for Integrated Assay	Reported Applications
Syringe pump	Simple	Syringe pumps	Need many tubing and syringe pumps for integration	Dose-response assay33,34 Embryonic stem cell culture13,35,36 Apply shear stress 37 Gene expression analysis 15 3D tissue culture38,39 High-content drug assay 17
Pneumatic pressure	Simple but need liquid reservoir	Pressure source and regulator	Need only one to a few connections to pressure source	Dose-response assay 40 Drug toxicity 32
Gravity force	Simple	Syringe or intravenous (IV) infusion set	Need many syringes or IV set for integration	Fibroblast culture 41 3D tissue culture 42
Capillary force	Simple	None or small absorbent	Need many surface areas or adsorbents for integration	Microarray culture 43 Fluid control and switching44,45
Surface tension	Relatively simple	Inkjet printing	No connection tube is needed	High-density culture 46 3D tissue culture47,48
Deformation of microchannel	Need multilayer microfluidic device	Pneumatic module or Braille module	Need software to control operating modules	Embryonic stem cell culture 49 Embryo culture 50

Perfusion Culture Microchamber Array Chip

Figure 1 shows our pressure-driven perfusion culture microchamber array chip with its simple interface. ³² The microchamber array chip has macroscopic liquid reservoirs, which can stock a small volume (<1 mL) of culture media containing different drugs or cell suspensions loaded by a micropipette ( Fig. 1a ). The culture media can be delivered simultaneously through the microfluidic network from the liquid reservoirs by applying pressure through the air vent filter ( Fig. 1b ). In addition, each microchamber was designed to have a unique geometry and independent perfusion microchannel to prevent cross-contamination between neighboring microchambers and for permitting easy air-bubble removal, uniform cell loading, and pressure-driven long-term perfusion culture in quantitative cell-based assays ( Fig. 1c ). We examined the cytotoxic effect of seven anticancer drugs in parallel to demonstrate the functionality of this pressure-driven perfusion culture microchamber array chip ( Fig. 1d ).

Figure 1.

Photographs of the pressure-driven perfusion culture microchamber array chip. (a) Stock of culture media containing eight different drugs by usual micropipette. (b) Loading of the culture media into the microfluidic network via a micropipette by applying pressure. (c) Loaded solution without cross-contamination or air bubbles in the microchamber array. (d) Fluometric cell growth measurements for the drug cytotoxicity assay. Reproduced from Sugiura et al. ³² with permission from John Wiley.

Microfluidic chips are usually fabricated from poly(dimethylsiloxane) (PDMS), which is a flexible, durable, transparent, and inexpensive polymer that allows microstructures to be easily molded by soft lithography.^52–54 In our previous study, we designed a PDMS microchip with a microfluidic network composed of microchannels at different depths by means of multilayer photolithography and replica molding. ⁸ The dimensions of the microchamber and connected microchannels were determined from calculations based on mass transfer and hydrodynamics. Table 2 shows the parameters to be considered for designing dimensions of the microchambers and microchannels. These calculations were used to predetermine diffusive and convective mass transfer rates and velocity distribution at medium flow (Fig. 2), allow the microfluidic network to load a specific amount of cells into the microchamber array, and control the supply of oxygen, nutrients, and soluble factors (e.g., growth factors and cytokines). Calculations based on mass transfer and hydrodynamics are effective in designing the cell culture microfluidic chips with a controlled microenvironment within the microchamber.

Table 2.

Parameters to Be Considered for Designing Dimensions of the Microchambers and Microchannels.

	Unit	Equation	Physicochemical Meaning
Medium exchange rate	µm3/cells/day	Given parameter	Nutrition supply
Pressure drop	Pa	Given parameter	Applied pressure
Shear stress	Pa	6 µQ/wh2a	Physical signaling and damage to cells
Reynolds number	—	ρUDeq/µ b	Inertial force/viscous force
Peclet number	—	LU/D c	Convection/diffusion

a

µ denotes the fluid viscosity, Q denotes the flow rate of the fluid, and w and h denote the channel width and height, respectively.

b

ρ denotes the fluid density and U denotes the fluid velocity. D_eq denotes the equivalent diameter and was calculated as D_eq = 2wh/(w + h).

c

L is the diffusion length and D is the diffusion coefficient.

Figure 2.

Design of the microchamber array. (a) Enlargement of the microchamber. (b) Flow velocity distributions in the microchamber. Reproduced from Sugiura et al. ³² with permission from John Wiley.

Dose-Response Assay by Perfusion Culture Microchamber Array Chip

Dose-response assays are crucial in the field of drug discovery. Scientists often prepare stepwise drug concentrations in a linear^55–59 and logarithmic^60–65 dilution series and evaluate the 50% growth inhibitory concentration (IC₅₀) or 50% effective concentration (EC₅₀). Generally, the dose-response assays are carried out in multiwell plates and performed either manually or semiautomatically with expensive robotics.^66,67 The preparation of a large number of drug solutions at different concentrations is, however, tedious and time-consuming. In addition, the dilution process sometimes causes experimental errors, since serial dilution using a micropipette can induce accumulation of dilution errors. To address these issues, serial dilution microfluidic networks for generating various concentration profiles have been developed.^34,68–72

We previously described a serial dilution microfluidic network for generating arbitrary monotonic concentration profiles (Fig. 3). ⁵¹ The serial dilution microfluidic network was composed of microchannels with different fluidic resistances, extremely thin fluidic-resistance microchannels, and thick diffusion-mixing microchannels and was designed using calculations based on mass transfer and hydrodynamics. In the serial dilution microfluidic network, high-concentration and low-concentration solutions were continuously introduced into the inlet microchannels. The high-concentration solution was then diluted with the low-concentration solution in the diffusion-mixing microchannel located at the meeting point of the two solutions ( Fig. 3a ). The dilution ratios were determined by the flow rates of the two solutions flowing into the diffusion-mixing microchannels. Finally, the diluted solutions flowed into each outlet port according to predetermined monotonically decreasing functions. We describe here a general procedure to design the microfluidic network:

1. Determine a microchamber size and cell density considering the target application.

2. Determine a desired concentration profile considering the target application.

3. Determine a possible applied pressure range considering available equipment.

4. Estimate flow rate in each outlet microchannel from the value estimated by step 1.

5. Calculate flow rate, Q, in each microchannel considering the splitting and mixing ratio estimated by the concentration profile determined in step 2 and the flow rate at the outlet estimated by step 4.

6. Calculate fluidic resistances, R, of each microchannel considering the applied pressure, ΔP, determined in step 3 and Q in each microchannel calculated in step 5. These calculations can be performed by the following equation based on analogy to an electric circuit ⁷³ :

   $$\Delta P=R\times Q$$
   .

7. Finally, calculate dimensions of each microchannel assuming the steady-state pressure-driven flow in a rectangular microchannel formulated as

$$R=\frac{12\mu l}{w{h}^3}{\left\{1-\frac{h}{w}\left[\frac{192}{{\pi }^5}\sum _{i=1,3,5}^{\infty }\frac{1}{i^5}\tanh \left(\frac{i\pi w}{2h}\right)\right]\right\}}^{-1}$$

where l is the channel length, w is the channel width, h is the channel height, i is the index variable for sigma, and µ is the fluid viscosity. ⁷⁴

Figure 3.

Automatic generation of arbitrary concentration profiles by the serial dilution microfluidic network. (a) Composition of the serial dilution microfluidic network. (b) Generated logarithmic concentration profiles. Reproduced from Hattori et al. ⁵¹ with permission from The Royal Society of Chemistry.

The microfluidic networks of this system were composed of fluidic-resistance microchannels with a cross-sectional area approximately one-tenth that of other microchannels. The fluidic resistance in these microchannels per unit length is therefore 10²- to 10³-fold higher than that of other microchannels. The flow rates for each diffusion-mixing microchannel are determined by the fluidic-resistance microchannels. Therefore, arbitrary concentration profiles can be generated by constructing fluidic-resistance microchannels at an appropriate length and ensuring complete mixing in the diffusion-mixing microchannels ( Fig. 3b ).

The optimization of the mixing process is essential for the accurate performance of the serial dilution microfluidic network. In these chips, the high-concentration and low-concentration solutions are mixed with each other by diffusion in the diffusion-mixing microchannels. The mixing process can be examined by simulating the flow pattern and concentration distribution in the diffusion-mixing microchannels by commercial hydrodynamic analysis software ( Fig. 4a ; we used version 3.3a; COMSOL Multiphysics, Stockholm, Sweden). The required residence time for complete mixing in the diffusion-mixing microchannels depends on the molecular weight of the solute and the dimensions of the diffusion-mixing microchannels. We can evaluate the required residence time by calculating the dimensionless diffusion time, Fick number (F_i), expressed as the following equation:

$$F_i=\frac{Dt}{r^2}$$

where D is the diffusion coefficient, t is residence time, and r is the characteristic length defined as half the width of the diffusion-mixing microchannel. We can use the calculated F_i value to design the structure of a diffusion-mixing microchannel suitable for the molecular weight of the solute and the flow rate in the serial dilution microfluidic network ( Fig. 4b ). Our design, based on mass transfer and hydrodynamics, enabled the serial dilution microfluidic network to be sufficiently compact for integration into the perfusion culture microchamber array chip, to generate stable and accurate concentration profiles for several days over a wide range of flow rates.

Figure 4.

Predetermination of flow and mixing in the microchannel. (a) Flow pattern and concentration distribution in the diffusion-mixing microchannel by hydrodynamic simulation. (b) Average difference from the expected dilution ratio and dimensionless diffusion time by using the hydrodynamic calculation. Reproduced from Hattori et al. ⁵¹ with permission from The Royal Society of Chemistry.

Recently, we developed microfluidic cell-based assays for IC₅₀ determination by applying the pressure-driven perfusion culture microchamber technology and the serial dilution microfluidic network ( Fig. 5a ). ⁴⁰ In these experiments, the perfusion culture microchamber array chip was equipped with 12 perfusion culture microchambers and a serial dilution microfluidic network capable of generating 12 different logarithmic concentrations spanning six orders of magnitude ( Fig. 5b ). We successfully used this pressure-driven microfluidic perfusion culture device to perform a dose-response assay for paclitaxel in HeLa cells by fluorescent image analysis. We successfully calculated the IC₅₀ of paclitaxel using data from three separate microchamber arrays ( Fig. 5c ).

Figure 5.

Photographs of the perfusion culture microchamber array chip equipped with a serial dilution microfluidic network. (a) Cell culture microfluidic chip. (b) Enlargement of the serial dilution microfluidic network and the cell culture microchambers. (c) Fluometric cell growth measurement for the drug cytotoxicity assay. Reprinted with permission from Sugiura et al. ⁴⁰ Copyright 2010, American Chemical Society.

Integrated Perfusion Culture Microchamber Array Chip

More recently, we developed a microplate-sized integrated perfusion culture microchamber array chip (Fig. 6). ⁷⁵ The culture media contained 12 different drugs, and the dilution solution (normal culture media) and cell suspension were injected via the inlet or outlet ports and could be stocked in the macroscopic reservoir by using a micropipette ( Fig. 6a ). The integrated perfusion culture microchamber array chip was equipped with an array of 384 microchambers arranged according to the Society for Laboratory Automation and Screening (SLAS) standard microplate format ( Fig. 6b ). Microchamber arrays complying with the SLAS standard format can be scanned by commercial microplate readers.

Figure 6.

Microplate-sized integrated perfusion culture microchamber array chip. (a) Photographs of the integrated perfusion culture microchamber array chip. (b) Enlargement of a microchamber array unit. Reproduced from Hattori et al. ⁷⁵ with permission from CBMS.

The serial dilution microfluidic network described here has a similar structure to that described in our previous study in that the drug solution was diluted at a ratio of 10^0.5 for each dilution step to create a concentration profile spanning three orders of magnitude over six dilution steps. ⁷⁵ By using the integrated perfusion culture microchamber array chip, parallel quadruplicate assays for 12 drugs at eight different concentrations, which is a total of 384 assays, can be performed simultaneously.

We also developed two devices, a dispenser unit and a perfusion culture unit, for automation of the integrated perfusion culture microchamber array chip (Fig. 7). Both devices are designed to apply the appropriate pressure to the macroscopic reservoir to induce flow into the microfluidic network. Figure 7a shows the dispenser unit, which can load cells into each microchamber, transfer liquid into the microchannel, and be used to rinse the liquid reservoirs. Figure 7b shows the perfusion culture unit, which can be used to maintain culture conditions (flow rate of media, CO₂ concentration, humidity, and temperature) during assays. These semiautomatic perfusion culture devices have the potential to increase the efficiency of microfluidic devices for high-throughput cell-based assays.

Figure 7.

Peripheral devices for the automation of the cell-based assay with the integrated perfusion culture microchamber array chip. (a) Dispenser unit. (b) Perfusion culture unit.

In conclusion, we describe an efficient pressure-driven perfusion culture microchamber array chip with a user-friendly interface and well-designed microfluidic network. Cell suspension and culture media containing different drugs can be loaded with a micropipette and can be delivered simultaneously into the microfluidic network by applying pressure through the air vent filter. We have also presented the means to design the microfluidic chips. The pressure-driven integrated perfusion culture microchamber array chip is a versatile system that can be used with standard microplate readers and dispensing robotics and has the potential to be an efficient platform for a microfluidic cell-based assay in drug discovery.