Microfluidic flow cells: off-chip pH monitoring for organ-on-chip

Microfluidic Cell Culture: a way toward flow cells

Nevertheless, it is a reductionist approach that fails to represent the intrinsic complexity of cell differentiation, and interaction, tissue function, drug response, and the dynamic and intricate microenvironment present in in vivo biology (Fig. 1.). These physical and chemical signals are crucial for cellular development and behaviour, and their absence leads to poor translation between in vitro and in vivo systems. Furthermore, traditional cell culture protocols are time-consuming and labour-intensive].³ These limitations have profound effects on the health and pharmaceutical industries, resulting in inefficient and increasingly costly drug development processes. It is estimated that roughly 40% of new drug candidates that reach clinical trials fail due to unforeseen toxicity effects in humans.⁴ Closing the gap between preclinical and clinical data translation calls for better methods to predict human responses.

Figure 1. Illustration of the signals in the cell microenvironment. Chemical and temperature gradients, cell-to-cell interactions, molecular signalling from different cell types, and physical stimuli are vital for cell development and behaviour, being poorly represented in traditional cell culture. Reproduced from Coluccio et al.¹

Miniaturising cell cultures into microfluidic systems can address some of the mentioned drawbacks and can also provide other interesting features. The minute volumes dealt with in microfluidics substantially reduce the necessary amount of reagents required for traditional cell culture, which, besides directly reducing costs, is especially advantageous when handling limited or rare samples, such as patient primary cells (those collected directly from patients as opposed to immortalised cell lines⁵). Also, handling smaller quantities of cells decreases the heterogeneity of the culture, so complex cell behaviour is more readily distinguishable. The easiness of fabrication of microfluidics devices provided the foreground for designed co-culture chips, in which the key aspects of the cellular microenvironment can be reproduced, i.e. the co-culture of different types of cells without cross-contamination, and replication of physical stimulus and chemical gradients with high spatiotemporal precision due to user-controlled flow.³ The advent of flowing media continuously or on-demand on top of the cells is of particular interest because it not only replicates the dynamic environment of the human body, it also automates much of the laborious cell culture practices. That, allied with the high level of control provided by the laminar flow and predefined channel geometries, leverage the precise addition of compounds of interest or sampling of the microenvironment without much disruption to the cells. Moreover, high throughput can be achieved by parallelising multiple microfluidic devices, in which in situ monitoring and analysis can be integrated and performed in a continuous manner.^1,3

The advantages of microfluidic cell culture, which can easily mimic the cellular microenvironment and provide a high level of control in addition to automated monitoring and analysis, led to the development of the “organ-on-chip” (OoC) field in the 2010s, currently one of the most relevant areas of research in microfluidics.⁶ OoCs are specialised in vitro cell culture models based on microfluidic devices that are structured around three main pillars: reproducing the mechanical and biochemical stimulus of the cellular microenvironment; simulating the 3D microarchitecture of tissues with multiple cell types placed in specific communicating compartments; and, replicating functional tissue-tissue interfaces via cell-cell interaction (Fig. 2.a.). The main goal is to reproduce complex in vivo biology to better understand cellular behaviour and metabolism, disease phenotypes, and drug response without the need for insufficient in vitro models or imprecise animal models.⁷

Several organ-on-chip devices have been successfully developed, namely, liver⁸, lung^7,9, kidney¹⁰, heart^11,12, intestine¹³, bone¹⁴ and bone marrow¹⁵, nerve¹⁶, blood vessels^17,18, and blood-brain barrier¹⁹ (Fig. 2.b.). They have demonstrated drug responses closer to human physiology when compared to standard in vitro models, with the enhanced complexity leading to better cell differentiation and drug transport.²⁰ As OoC models mature, the interconnection of several OoCs in a “human-on-chip” approach becomes feasible and has been investigated by several groups in different configurations.^21–24 An emerging area that has the potential to deeply change modern medicine is the development of OoC with induced pluripotent stem cells derived from patients. This has the potential to be a paradigm shift in personalised medicine and the development of patient-specific treatments.

Figure 2. Organ-on-chip systems. a. Schematic of the cross-section of a specialized organ-on-chip device, a lung-on-a-chip, highlighting the tissue interface and the vacuum chambers for mechanical stimulus. Adapted from Huh et al.⁷. b. Illustration of different organ-on-chips, specifically, bone-, liver-, gut-, lung-, heart- and kidney-on-chip.

The increase in complexity of cell cultures enabled by microfluidics opened the possibility of more in-depth spatiotemporal analysis of cellular development and behaviour, further increasing the importance of tools such as live-cell imaging²⁵, metabolic detection and analysis²⁶, and microenvironmental monitoring of parameters such as pH ²⁷. Consequently, the long-term culturing of cells in OoCs moved from the CO₂ incubator to the microscope stage. As the complexity of the systems increased and the access to more detailed information deepened, enhanced resolution and real-time monitoring of the systems became a clear demand ^26,28. Thus, a growing need to more closely monitor temperature, pH and dissolved O₂, allied with increased interest in having a better understanding of the in situ metabolic activity, led to a plethora of on-chip and off-chip sensing solutions.

The importance of pH monitoring

As previously mentioned, the gap between preclinical and clinical data causes huge losses for the health sector, besides delaying access to treatment for patients. One of the contributing factors of this global issue has been attributed to the variables of microenvironment conditions. ²⁹ Cell lines require a stable pH to grow, as acidic pH can irreversibly inhibit enzyme activities and the synthesis of DNA, RNA, and protein, thus compromising cell viability. ^30,31 pH is defined as the concentration of free H⁺ ions in a solution, also called the activity of protons, and can be calculated by the Henderson-Hasselbalch equation:

in which pk_a is the acid dissociation constant; [HB] is the concentration of a compound’s protonated form; and [B] is the concentration of a compound’s unprotonated form.

Buffering systems are capable of maintaining the pH constant due to a dynamic equilibrium. The equilibrium between the protonated and unprotonated forms of the buffer can move towards reacting to new free H⁺ or releasing H⁺ ions to compensate for ions that were added or removed from the surrounding solution. Both mechanisms have the effect of keeping the previous free H⁺ concentration constant to a certain extent. If the addition or removal of free H⁺ goes beyond a threshold that can be absorbed by the buffer, the pH will change, i.e. the buffering capacity of the solution was surpassed³². That is why complex organisms, such as mammalians, have developed dynamic buffering systems that adjust the concentration of the buffer according to the system’s needs. The physiological buffering system is based on the equilibrium of CO₂/HCO₃^– (bicarbonate buffering system, pk_a= 6.15). The lungs, through the gas exchange of CO₂, and the kidneys, through ion transport proteins, are responsible for keeping the ratio of the protonated to unprotonated species of this buffering system in homeostasis, and the pH of the organism within the physiological range (Fig. 3.a.). Cell metabolism is bound to acidify the pH of the microenvironment due to the production and release of lactate, which reacts with water to form acid lactic. For this reason, cell culture media usually contains a buffering system to keep pH under physiological conditions.³³ Besides the CO₂/HCO₃^– buffer, media can be buffered by non-volatile buffers (NVB), such as HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (pKa = 7.3; 37 °C), PIPES (piperazine-N,N′-bis(2-ethanesulphonic acid); pKa = 6.7) and MES (2-(N-morpholino)-ethanesulfonic acid; pKa = 6.0)³⁴, however, NVBs in cell culture media is out of the scope of this review, detailed work on the topic can be found in Michl et al., 2019.³⁵

The bicarbonate buffering system is replicated in cell biology labs with the aid of a CO₂ incubator. This system works through the equilibrium of CO₂ from the CO₂-rich atmosphere of the incubator, which dissolves in the media and reacts with H₂O forming carbonic acid, and the NaHCO₃ present in the media (Fig. 3.b.). Cell culture media come with different concentrations of NaHCO₃ and require different percentages of CO₂. For example, DMEM comes with 44 mM of NaHCO₃, which requires approximately 10% of CO₂ in the atmosphere to maintain the pH close to 7.4. CO₂ incubators are conventionally set to 5% CO₂, which keeps the pH of DMEM around 7.6-7.8. The production of lactic acid and CO₂ by healthy cells and the buffering capacity of serum, often added to DMEM, partially offsets this difference, allowing the growth of cells with DMEM using conventional CO₂ concentrations.³⁵ However, the implications of this adjustment are not well-understood. The intracellular pH of cells is in close equilibrium with the extracellular pH of the microenvironment due to the transmembrane ion transporter proteins. Michl. et al.³⁵ studied the effect of extracellular pH change on intracellular pH and concluded that cells re-balance the internal pH based on the external one. Since most H⁺ targets are intracellular, this change can affect intracellular mechanisms and metabolic pathways in unpredictable ways.

Figure 3. Bicarbonate buffering system. a. Illustration of the bicarbonate buffering system in the human body, highlighting the equilibrium provided by lungs and kidneys. b. Bicarbonate buffering system inside the CO₂-rich atmosphere of the CO₂ incubator reacting with the salt present in the media.

When the culturing system is downscaled to micron-range in OoCs, substantially reducing the volumes, the monitoring of pH becomes even more important. Small variations in pH, e.g. 7.4±0.3, are tolerated in traditional cell culture; however, the more accentuated diffusional processes of microfluidic cell culture result in a more pronounced effect on cell viability, prompting closer monitoring of crucial environmental parameters.³¹

Monitoring cell culture environment with flow cells

The measurement of metabolic activity in microfluidic cell culture is usually done with on-chip solutions. For example, Weltin et al.³⁶ developed a microphysiometer system in a microfluidic chip with dissolved O₂ and pH electrochemical sensors inside the culturing chamber. The authors successfully cultured brain cancer cells with online and continuous measurement of the cells’ metabolism. However, these solutions prove too complex and costly³⁷  when the goal is to monitor the cell culture environment, which does not require the same level of detailed information. For that, microfluidic off-chip solutions, such as microfluidic flow cells, are more apt. Microfluidic flow-through cells can be developed in-house ^38–40, which are usually 3-D printed⁴¹, or commercially available^42–46. The format can vary according to the application, such as T-junctions for HPLC applications or other geometrical structures fitted to house the sensor (Fig. 4). The main goal is to place the sensors in line with the microfluidic system to allow for continuous monitoring.

Figure 4. Cell culture pH monitoring. a. Examples of different flow-through cells for pH monitoring. The first three are commercially available from different suppliers, the fourth was developed and 3D-printed in-house. b. Schematic of a microfluidic setup with an in-line microfluidic flow cell and pH sensor.

A good example is the work of Zhang et al.⁴⁷ who developed a “physical sensing unit” composed of pH, O_2, and temperature sensors in a flow-through cell developed in-house to monitor the microenvironment of a complex organ-on-chip system. The system was designed as a modular platform to control two organ-on-chips connected to an automated flow control breadboard. Interestingly, the biomarker sensing unit, i.e. the sensors to measure the metabolism, was also off-chip. The platform was validated with a drug screening assay using liver and heart organ-on-chips, exposed to acetaminophen for 5 days and doxorubicin for 24 h, where cytotoxicity biomarkers were detected and quantified.

Farooqi et al.²⁶ 3-D printed a pH sensing flow-through cell to house an optical pH sensor built with commercially available parts that measured pH based on the colour of the phenol red present in the media. The goal was to monitor a live-on-chip system also intended to investigate the drug toxicity of doxorubicin. The researchers were able to detect a decrease in pH when cells were exposed to the high concentrations of the drug when compared to the control. The same group performed similar work with a lung-on-chip system⁴⁸, attesting to the robustness of the developed sensing unit. Fibroblasts had pH and O₂ monitored for 3 days in a similar fashion by Ali et al.⁴⁹ Also, using the phenol red of the media, Wu et al.⁵⁰ designed a high throughput pH sensing unit, which consists of parallelised microfluidic channels with a sensing chamber connected to optical pH sensors. They performed simulations to define the best shape of the sensing chamber, resulting in an oval shape, and the thickness of the PDMS layer has also been decreased to minimise the loss of light transmission. Optical sensors seem the most common choice for microfluidics flow-through cells due to their independence of reference electrodes, electrical connections and flow rates. Also, they are not prone to biofouling, corrosion and interference from electrochemical signals from molecules present in the medium.⁴⁹

Off-chip flow cell for monitoring Conclusion and Outlook

The important growth of organ-on-chip technologies brought with it an increased need to monitor more closely what was happening inside the chips. Several on-chip and off-chip solutions have been designed, and microfluidic flow cells seem to be the way forward for microenvironment pH monitoring. On-chip sensors are suited for the detection of metabolic activity that can benefit from high spatiotemporal resolution. However, these solutions are too costly and complex for microenvironment monitoring, which does not require the same level of precision. For microenvironment monitoring, microfluidic flow cells with embedded optical sensors, which are placed in the microfluidics circuit for real-time and continuous monitoring, have been the most common choice.

References