Microfluidics for Organ-on-chip Cell culture

Organ-on-a-Chip Applications and Models

In short: organ-on-a-chip applications fall into three families — therapy development, drug discovery and toxicity testing, and regenerative medicine. Each is realized through organ-specific models: lung, liver, gut, kidney, heart, tumor and blood vessel on chip. What separates a working model from a promising one is usually the stability of the perfusion driving it. 

Therapy development  

Organ-on-chip platforms have proven their flexibility and robustness in drug screening by borrowing materials and techniques from tissue engineering and microfabrication. The clearest trend in the literature is the use of human induced pluripotent stem cells (hiPSC) to build personalized organ models. Stem-cell research of this kind makes patient-specific screening practical rather than theoretical. 

Patient-derived cells behave, under perfusion, closer to the way they behave in the donor, which is what makes personalized screening credible rather than aspirational. Cell culture chips joining a culture well to a microfluidic channel through a porous membrane let users study complex configurations in a single device. 

That geometry is the standard choice for air–liquid interface (ALI) culture and for endothelium/epithelium barrier and crosstalk studies, where two channels are perfused independently and the membrane between them carries the readout. It is also the simplest way to observe interactions between two cell types kept in separate compartments. 

Barrier integrity is a flow-sensitive endpoint. A blood–brain barrier model reports physiological permeability only if its endothelial layer has been conditioned under a shear profile it recognizes, which makes a human BBB-on-chip permeability study as much a fluid-handling result as a cell biology one.

Drug discovery

Emerging in-vitro tissue culture platforms are useful chiefly for predicting the human response to new compounds. Microphysiological systems of this kind give pharmacology better tools for evaluating how drug candidates act on human tissue, earlier and with fewer species-translation assumptions than a conventional pipeline allows. 

Organ-on-chip models support accurate prediction and mechanistic investigation of dose-limiting human toxicity, and the exploration of approaches that mitigate the potential toxic effects observed. Applied early in drug development, they inform the identification, modification and optimization of lead compounds. 

The aim is safer molecules with a better chance of surviving clinical trials, and fewer failures at the preclinical-to-clinical boundary — the point at which animal data and human responses diverge most expensively. 

Two properties matter for screening. Throughput: multiple chips running under identical, independently controlled flow rather than one at a time. And dose fidelity — a compound has to reach the cells at the intended concentration for the intended duration, which is a question of injection volume and timing, not chip design. 

Regenerative medicine

Regenerative medicine uses organ-on-a-chip differently: not to test a compound against tissue, but to build and mature the tissue itself, a route with obvious therapeutic potential. Perfused culture sustains the differentiation of stem-cell-derived constructs across the weeks such protocols demand, and is now central to personalized medicine research, renewing nutrients and oxygen continuously and clearing waste before it accumulates. 

The same devices are used to study how mechanical and biochemical cues direct cell fate. Controlled shear, stretch and gradient exposure are inputs to differentiation rather than stressors to be survived, and setting them precisely is what turns a culture protocol into an experiment. Work on organoid culture in microbeads and on biochemical and biomechanical stimulation illustrates both routes. 

A third strand is graft and construct assessment: perfusing an engineered tissue to confirm that it survives, vascularizes and stays functional under flow before it goes near an animal model. Long-term recirculation is the enabling capability here, and the one most often improvised. 

Applications by organ: lung, liver, gut, kidney, cancer, vessel

In short: the models below have the deepest published record. Each carries a characteristic mechanical signature — a shear range, a stretch cycle, a gradient — that the perfusion system has to reproduce faithfully. 

Lung and airway. Airway models are built at an air–liquid interface: epithelium exposed to air on one face, medium perfused on the other. Alveolar models add cyclic stretch to mimic breathing. See our case study of Lung-on-chip. Both are demanding, because an ALI culture tolerates none of the pressure transients liquid-filled channels would absorb. The air–liquid interface and co-culture chip is the usual starting geometry.  

Liver. Hepatic models are the reference application for metabolism and toxicity testing, and the reason many groups adopt the technology at all: primary hepatocytes lose function quickly in static culture and stay functional far longer under perfusion. Because drug-induced liver injury is a common cause of late-stage failure, a liver model that remains metabolically competent for weeks has direct clinical value. See the CNRS/UTC liver-on-a-chip model. 

Gut and intestine. Intestinal models depend on shear to drive villus-like differentiation and barrier formation, so flow rate is an experimental variable rather than a setup detail. Fluigent published research on a human gut-on-chip assessing the effect of shear stress on intestinal function. Gut devices also host most microbiome and infectious disease research, where host–microbe interactions are followed under a stable low flow for days, with sampling for downstream analysis. 

Kidney. Proximal tubule and glomerular models reproduce filtration and reabsorption under defined tubular shear, and are used heavily for nephrotoxicity screening — the kidney concentrates what it filters, leaving it disproportionately exposed. Tubular epithelium polarizes and builds functional transporters in response to flow, so the perfusion profile is part of the model. See the biomechanics of a perfused kidney-on-chip model. 

Heart. Cardiac models are among the most developed, because human stem-cell-derived cardiomyocytes give a directly observable functional readout: beat rate and contraction. Heart-on-chip technology is used mainly for cardiotoxicity screening, where clinical relevance depends on using human cells, rather than inferring drug responses from animal data, and for modeling inherited cardiac disease. 

Cancer and tumor-on-chip. Tumor models study the microenvironment, invasion and drug responses, increasingly testing immunotherapies against patient-derived material. Interstitial flow and gradient control are central, since cell–matrix interactions and interstitial transport shape it. A chemotactic gradient held steady for days is what distinguishes a migration assay from an impression of one. See our review of cancer-on-chip models. 

Blood vessel-on-chip. Vascular models — artery, microvessel, blood–brain barrier — are where flow control is least optional, because endothelial phenotype is set by the shear waveform applied to it: endothelial cells align with the flow direction and remodel their junctions accordingly. Laminar, pulsatile and disturbed flow produce measurably different cells from the same endothelial monolayer. See the artery-on-a-chip model and the BBB permeability study. 

Beyond the seven. The technology extends to increasingly complex targets, and to tissue still under development in most labs. It reaches to endocrine and musculoskeletal tissue — a pancreas-on-a-chip study, cartilage-on-chip under complex mechanical stimulation, and skeletal muscle through the MYOCHIP H2020 project. A multi-organ platform connects several different compartments in one circuit, raising the requirement again: multiple independently regulated flows, held for weeks, without cross-contamination. 

New to the field? Start with our 5 tips for organ-on-chip models. 

Importance of fluid handling for organ-on-a-chip cell culture

In short: perfusion determines whether a model is physiological. Continuous flow renews nutrients and oxygen, applies the shear the tissue expects, and delivers biochemical information on a defined schedule. Its stability sets the reproducibility of everything measured downstream, which is why flow control systems belong in the experimental design rather than the plumbing. 

Thanks to a partnership with Beonchip, Fluigent offers cell culture devices matched to the application: to study cell culture under flow, to explore crosstalk between 2D and 3D cultures in biomimetic environments, to apply gradients to 3D cell cultures, and to join a culture well to a channel across a porous membrane. The full chip range sets out which design suits which model. More information on microfluidic cell culture chips 

Optimized cell culture activity. Constant perfusion renews nutrients and oxygen continuously, promoting growth and maintaining activity through long-term culture — the difference between a model that survives three days and one that still reports after three weeks. 

Biomechanics. The technology opened up the study of mechanical strain in cell biology by reproducing key aspects of the in-vivo microenvironment. Combining microfluidics with microfabrication and engineering techniques lets researchers recreate, at cell scale, the forces living tissue experiences — including the contraction cycles of heart muscle. 

Passive stimulation induced by shear flow. Liquid flow exerts shear stress on cultured cells and tissue, with substantial effects on growth, phenotype and gene expression. Shear is not a side effect to minimize; it is a parameter to set and hold. 

Active mechanical stimulation arises from organ function itself. Lungs, muscle and intestine are in constant motion, and their cells are subject to compression and stretch. Organ-on-a-chip culture reproduces those dynamic cycles under controlled conditions — see passive and active mechanical stimulation. 

Biochemical studies. Cells are exposed to biochemical stimulation from the embryonic stage through adult life, and the spatiotemporal regulation of those signals determines cell fate, phenotype, metabolic activity and pathological behavior. Reproducing that variation in vitro demands fast response and high stability from the instrument generating it. 

Flow control systems and Fluigent’s added value

In short: automated fluid delivery raises reproducibility by removing the variability of manual handling. Among the systems available, pressure-based control is the technology of choice for organ-on-chip, because it combines flow stability with fast response — the two properties long cultures and stimulation protocols depend on. 

Implementing perfusion and automated delivery brings compounding advantages over manual pipetting: reproducibility improves, hands-on time falls, and every parameter that matters — delivery time, volume, injection speed — is regulated rather than approximated. For a multi-week culture it is also the only practical way to hold conditions constant overnight. 

Automation changes what an experiment can contain. Once delivery is under software control, sequences that would be impractical by hand become routine: a stimulation protocol timed to the minute, sampling at fixed intervals for offline analysis, medium recirculated so that the factor content of the medium rises as it would in tissue, or an immunostaining sequence run overnight on a chip that never leaves the incubator. 

Several technologies address sub-millimeter fluid management. As demand has grown for microfluidic pumps offering higher stability, faster response and automation, pressure controllers have become the default. The Flow EZ is the standalone instrument in that family, usable without a computer and scalable into the LineUP series as a setup grows from a single channel to multiple channels. 

Response time and stability

These systems work by pressurizing the sample reservoir to control the pressure drop between the inlet and outlet of the microfluidic circuit. Because no moving part displaces the liquid, the generated flow is smooth by construction, and its responsiveness follows the responsiveness of the pressure controller itself. 

The practical consequence shows up in comparison. Peristaltic pumps deliver flow in pulses; cells register those pulses, and endothelial layers respond to them. Our comparison of peristaltic pumps and pressure-based flow control sets out the difference on the same model. 

Stability matters across weeks as well as milliseconds. A drift of a few percent is invisible on a screen and decisive in a three-week differentiation protocol, which is why long-run stability, not peak specification, is the number worth comparing between devices. More information on flow controllers