Triplicate Dynamic Cell Culture Platform for Enhanced Reproducibility in Anti-Cancer Drug Testing

Abstract

The development of stable and standardized in vitro cytotoxicity testing models is essential for drug discovery and personalized medicine. Microfluidic technologies, recognized for their small size, reduced reagent consumption, and control over experimental variables, have gained considerable attention. However, challenges associated with external pumps, particularly inconsistencies between individual pumping systems, have limited the real-world application of cancer-on-a-chip technology. This study introduces a novel triplicate cell culture system (Tri-CS) that simultaneously supports dynamic cultures in three independent units using a single peristaltic pump, ensuring consistent flow conditions. Our findings demonstrate that the Tri-CS significantly reduces variability compared to individual pump systems, enhancing the reliability of anticancer drug cytotoxicity testing. Furthermore, we evaluated gemcitabine cytotoxicity, which shows enhanced drug efficacy in dynamic conditions. Fluorescein diffusion tests revealed greater diffusion efficiency in dynamic cultures, which contributed to the higher observed drug efficacy. The potential for broader application of the Tri-CS, including its compatibility with commercially available transwells and the opportunity for use in more complex cancer-on-chip models, positions this system as a valuable tool for advancing microphysiological systems in preclinical research.

Keywords: drug cytotoxicity testing; dynamic cell culture; microfluidics; organ-on-a-chip; reproducibility.

Figure 1

Structural diagram and operating principle of the Tri-CS system. (A) The overall composition and a photo of the Tri-CS system, which consists of a culture chip, a peristaltic pump, and an external water reservoir, all connected to form a closed-loop system. (B) The detailed structure of the culture chip in the Tri-CS system, vertically divided into three layers: the micro pump, the micro valve, and the culture chamber. The water pool in the bottom layer of the micro pump is shared by the three units. The pink-shaded area represents the region where the medium is present. Scale bar = 1 mm (C) The micro valve consists of an O-ring and a PDMS membrane with openings. The design ensures unidirectional fluid flow. (D) Schematic of a single unit and operating modes of the culture chip. A unit is composed of a micro pump connected to the shared water pool, two micro valves, and a culture chamber. The pink arrows indicate the direction of medium flow. In Mode 1, the PDMS membrane in the micro pump moves downward due to lower pressure in the water layer, driving the medium above to flow in through the left valve. In Mode 2, the PDMS membrane in the micro pump moves upward due to higher pressure in the water layer, pushing the medium out through the right valve. Both micro valves contain PDMS membranes with openings, arranged in conjunction with an O-ring to function as unidirectional valves.

Figure 2

Medium flow stability between units in the Tri-CS system. (A) Working concept of a single unit. The water cycle is maintained via silicone tubing, allowing medium circulation within the device. The two different arrow colors indicate the direction of fluid flow in two distinct modes. External pump speeds ranging from 20 to 60 rpm result in internal medium flow rates between 458 and 1237 μL/min. (B) Flow patterns of the three units at external peristaltic pump speeds of 20, 30, and 60 rpm. Units are represented from top to bottom in blue, red, and green, corresponding to Unit 1, Unit 2, and Unit 3, respectively. (C) Average flow rates of the three units over 5 days at the three different flow speeds. Error bars represent the standard deviation (SD), with n = 3. (D) CV calculated for the three units over 5 days at different flow speeds.

Figure 3

Medium flow rate stability comparison between Tri-CS and individual pump Systems (A) Schematic of three units driven by the Tri-CS system. (B) Schematic of three units driven by individual pumps. (C) Table of adjusted pump speeds for individual pumps to achieve similar medium flow rates as in the Tri-CS system. (D) Comparison of the relationship between pump speed and medium flow rate across different units in the Tri-CS system, with each unit represented by a different color. (E) Comparison of the relationship between pump speed and medium flow rate across different pumps, with each pump represented by a different color. (F) Average flow rate and CV of different units in the Tri-CS system at five pump speeds (20, 30, 40, 50, 60 rpm). (G) Average flow rate and CV of different pumps at five pump speeds (3, 5, 7, 9, 11 rpm). (H) Comparison of the CV between the two methods, showing that the CV for Tri-CS averaged 3.5% across five pump speeds, while the average CV for the individual pump system was 8.4%, which was significantly higher than that of Tri-CS. P-values less than 0.001 are indicated by three asterisks (***). Error bars represent the SD, with n = 3.

Figure 4

Microscopy images and statistical data of cell viability in the Tri-CS system. (A) Bright-field images and live/dead fluorescent staining on day 3 for A549 cells cultured dynamically in the Tri-CS, with bright-field images taken on days 1, 2, and 3. Scale bar = 100 μm. (B) Statistical analysis of cell viability across three units (n = 5). Error bars represent the SD, with n = 5.

Figure 5

Drug cytotoxicity and diffusion testing in static culture, Tri-CS, and Single-CS systems. (A) Cell viability after treatment with 30 μM gemcitabine for 2 days across all groups. Black asterisks indicate statistically significant differences between groups, while red markers represent independent statistical results within each group. (B) CV values from the drug testing results for each group. (C) The concentration of fluorescence diffusion into the transwell was measured, with data collected at 4, 8, 12, and 24 h. Fluorescent concentration percentages were calculated based on a predetermined standard curve. (D) The CV values from the diffusion testing results showed a significantly higher variation in the Single-CS group compared to the other groups. P-values less than 0.001 are indicated by three asterisks (***), and “ns” indicates not significant. Error bars represent the SD, with n = 3.