Technologic advances change our lifestyle in many ways. New electronic devices become ever smaller but with ever increasing functions, so they can replace large space-occupying equipment. A few decades ago, microfluidics was developed using advances in the semiconductor industry and the subsequent expansion of microelectomechanical system field [1]. The technology can manipulate at the microscale and control fluid flow precisely. Microfluidics has provided new capabilities to researchers in biology and medical science. Since early 2000, published articles and patents on microfluidics have increased dramatically in the field of engineering. Importantly, although still a smaller overall fraction, articles on the use of microfluidics in the medical field are growing most rapidly at an exponential rate in recent years [2].

Microfluidics is characterized by the engineered manipulation of fluids at the submillimeter scale [2]. The microscale devices in microfluidics are commonly referred to as miniaturized total analysis systems or lab-on-a-chip technologies [2]. In the 1990s to 2000s, microfluidics focused on the miniaturization of previously macroscale assays. Within several years, various new technologies showed better performances than the traditional methods. For example, there are paper-based analytical devices for diagnostics [3], immunoassays assisted by surface tension [4], and organ-on-a-chip technologies [2,5] (Fig. 1). We will further focus on organs-on-chips.

Organ-on-a-chip is a microfluidic cell culture device that is a more physiologically relevant in vitro model than cells cultured in dishes. Important features include continuously perfused chambers inhabited by living cells arranged to simulate tissue- and organ-level physiology [5] (Fig. 2). So far, several individual organ-on-chips have been developed, including gut-on-a-chip [6], lung-on-a-chip [7], blood vessel-on-a-chip [8], cancer-on-a-chip [9], bone marrow-on-a-chip [10], and kidney-on-a-chip [11]. In addition, multiple organs-on-a-chip has been tried, such as liver-tumor-bone marrow-on-a-chip [12] or liver-skin-intestine-kidney-on-a-chip [13].

As shown in Fig. 3, the whole experimental process using organ-on-a-chip technology consists of several steps. Photolithography, a fabricating technique that can make channels in a microscale, comes from semiconductor technology. It requires labor-intensive work, clean environments, and sophisticated training. An organ-specific microenvironment is a basic platform in the design of these devices. Briefly, researchers design the optimal condition with the proper size, numbers, and compartments of the channels. According to these designs, fine molds are cultivated and the final devices can be made out of transparent and soft polydimethylsiloxane (PDMS) [5]. Sometimes, simple ways to make replicas can be used such as xerographic printing or three-dimensional (3D) printing of molds.

Cell sourcing is an important issue in the organs-on-a-chip field. Usually, a single cell type may be enough to mimic minimal organ function. For example, a kidney-on-a-chip for the evaluation of direct cellular toxicity may be not enough when constructed with renal tubular cells alone [11,14]. To mimic more physiologic conditions or evaluate cell-to-cell interaction, two cell types should be used like lung-on-a-chip [7], which has cultured endothelial cells on the bottom channels and alveolar cells on the top channels. This lung-on-a-chip also introduced human neutrophils with culture media to the bottom channels to observe the movement of inflammatory cells. Sometimes, cells in a tissue construct itself may be introduced into the devices, such as skin tissue or intestine constructs [13]. Humans-on-a-chip has multiple compartments to allow evaluation of the interaction among interesting organs [13].

The initial design of a published kidney-on-a-chip has two compartments [11,14]. A top channel mimics the urinary lumen and has fluid flow, whereas the bottom chamber mimics interstitial space and is filled with media. Kidney cells are under much lower shear stress than the endothelial or lung cells. This device used rat distal tubular cells or MDCK cells, and its shear stress was ~1 dyn/cm² [11]. A second report showed the same design, but human proximal tubular cells were used. In this model, the authors tried to reproduce cisplatin nephrotoxicity. Proximal tubular cells have much lower shear stress ~0.2 dyn/cm² [14]. The foot processes of podocyte, a glomerular visceral epithelial cell, form a size- and charge-selective barrier to plasma protein, and derangement of the barrier causes podocyte injury and proteinuria [17]. Some scientists have tried podocyte-on-a-chip, but it is still challenging. It may be because podocyte is exposed under a very low shear stress in vivo and requires a sophisticated culturing condition [18].

Kidney-on-a-chip has various potential applications (Fig. 4). Cisplatin nephrotoxicity was evaluated in the kidney-on-a-chip using human proximal tubular cells [14]. Cisplatin was introduced into the bottom space, and cisplatin-induced cellular damage was monitored to the cells for 24 hours. During the following 72 hours, shear stress was helpful for facilitating recovery of the injured cells and associated biomarkers. Shear stress in the devices can facilitate translocation of aquaporin-2 and relocation of actin cytoskeleton in the kidney-on-a-chip using primary cultured inner medullary collecting duct cells of rat kidneys [19], which is a good example of a physiological experiment using kidney-on-a-chip. Moreover, renal tubular epithelial cells are continuously exposed to the changes of extracellular microenvironment, e.g., transepithelial osmotic gradient, and changes of luminal or interstitial pH. The effects of these physiological factors on the functions of renal tubular cells could also be investigated by exploiting microfluidics [20,21]. If the offending antigens or mechanisms of the diseases are well identified, disease modeling can be easily fabricated in the kidney-on-a-chip. Kidney injury models using nephrotoxic drugs may be simple candidates for the application. Kidney-on-a-chip is only one compartment of an eventual multiorgans-on-a-chip system [13]. To research drug metabolism, renal excretion or metabolism should not be omitted. However, some multiorgans-on-a-chip systems have been using kidney tubular cells not for evaluating renal excretion, but for measuring kidney injury. Animal renal clearance is usually higher than human renal clearance. Using pharmacokinetic data from animal models runs the risk of underestimating human nephrotoxicity. Applications of human kidney-on-a-chip may fill such gaps through control of drug concentrations and flow rates that mimic human drug clearance or metabolism.

PDMS can absorb small organic compounds. Lipid soluble compounds, such as hydrophobic drugs, are incompatible with PDMS devices because of absorption into PDMS. The same extracellular matrix coated on PDMS can have different effects when coated on different surfaces such as glass or polystyrene or in the body in a cell type–dependent manner, requiring much trial and error. There is potential need for a universal blood substitute to perfuse multiple organ compartments. Some studies used a mixture of various cell culture media. Other studies used different culture media for each compartment. Another issue is scaling. Isometric shrinkage of organs creates an unbalanced system between organs because of functions that depend on organ volume versus organ area. There are multiple ways to scale down each organ, and what is most appropriate is still controversial [22,23]. Automated control systems can be helpful and may allow nonspecialists or academic researchers to replace 2D or 3D culture systems. Some companies have already launched commercial use of organs-on-a-chip. Pharmaceutical and biotechnology industries have a significant need for validation and prioritization of lead drug candidates. In addition, clinical trials may be able to adopt new biomarkers of drug efficacy, toxicity, or disease response developed using these devices in the near future.

Microfluidics is still a young and developing field. However, there are many opportunities for biologists and medical scientists to collaborate with engineers. For example, 3D organoid cultures, dynamic cultures, and disease modeling are areas where collaborations have led to better understanding of pathophysiology and mechanisms. Multiorgan-on-a-chip or human-on-a-chip is a young but exciting area. In the nephrology field, glomeruli-on-a-chip, kidney organoid, and nephron-on-a-chip are still challenging. Organs in the human body are vital but difficult to manipulate and image at the cellular level, making evaluation of their functions and responses difficult. In contrast, the use of organs-on-a-chip is relatively cheap, readily imaged, and easily accessible for measurement of functions and will provide a vital window into human physiology in ways not currently possible.

The authors have no conflict of interest.