What Is Organ-on-a-Chip? Benefits & Uses

What Is Organ-on-a-Chip?

An organ-on-a-chip is a microfluidic cell culture device with hollow channels lined with living human cells, continuously perfused with medium to replicate the key functions of a tissue or organ. It does not rebuild an entire organ. It reconstructs the minimal functional unit — an interface, a barrier, or a flow — that governs the behavior a researcher wants to study.

The defining feature is dynamic culture. Medium flows continuously through the channels, delivering nutrients, removing metabolic waste, and applying shear stress to the cell layer. This mechanical stimulation is the essential point: endothelial cells align with the flow direction and remodel their junctions, intestinal cells fold into villi-like structures, and hepatocytes retain metabolic functions they lose within days under static conditions. Cells read their mechanical environment as directly as they read a chemical signal, and a flat, still well plate sends them the wrong one.

The device’s scale is what makes this controllable. Channels measured in micrometres hold volumes of a few microlitres, where flow stays laminar and predictable. Under laminar conditions, a chemical gradient can be held stable across a channel for days, two fluids can run side by side without mixing, and shear stress can be calculated from geometry and flow rate rather than estimated. That predictability is why microfluidic engineering, and not simply smaller wells, underpins the field.

The technology is also referred to as microphysiological systems, a term favoured in regulatory documents because it describes the intent more precisely — recreating physiological conditions at microscale rather than miniaturising an organ. Both names cover the same class of devices, and both appear throughout the literature, so a search on either term returns the same body of work. A chip is a platform, not a finished model. What turns a device into a liver, gut, or blood-brain barrier model is the cell source, the extracellular matrix, the co-culture design, and the flow regime applied to it. The same chip geometry can host very different biology, which is why published protocols specify culture conditions in far more detail than device dimensions.

Key Components of a Microfluidic Culture Device

Four elements define any organ chip: the chip body and its channels, the cells, the extracellular matrix, and the fluidic control that drives perfusion. Each one is an experimental variable, and each affects the physiological relevance of the result as much as the biology seeded into the device.

The chip and its channels. Most designs use one or two microchannels, typically tens to hundreds of micrometres deep. In a two-channel format, the compartments are separated by a thin porous membrane, letting an epithelial layer sit on one side and an endothelial layer on the other, in contact but not mixed.

The cells. Primary human cells, immortalised lines such as Caco-2 or HUVEC, and human induced pluripotent stem cells are all used, and the choice is a trade-off between fidelity and practicality. Immortalised lines are robust, cheap, and reproducible but carry genetic abnormalities. Primary cells behave more like the source tissue yet vary between donors and survive limited passages. Stem cell sources are increasingly preferred because they allow patient-specific models and can be differentiated into several lineages from a single donor, letting one genetic background populate every compartment of a co-culture.

The matrix. Cells are seeded onto a protein coating or embedded within a hydrogel that supplies mechanical support and biochemical cues. Collagen, fibrin, and basement membrane extracts are common choices, and matrix stiffness measurably influences cell phenotype — the same cells behave differently on a soft gel and a stiff one. Where a porous membrane separates two channels, its thickness and pore size govern how freely cells signal to each other across the interface.

The fluidic control. Perfusion is what separates a chip from a well plate, and it is the component researchers most often underestimate. Flow rate sets shear stress, pressure stability determines whether that stress is constant or fluctuating, and the ability to recirculate a small medium volume for days or weeks decides whether a long culture is feasible at all. Bubble management, sterile connections, and the dead volume of the circuit are equally practical constraints: an air bubble crossing a channel can strip a monolayer in seconds and end an experiment that has run for a fortnight.

Figure 1: Cross-section of a dual-channel organ-on-a-chip device. An epithelial layer and an endothelial layer are cultured on either side of a porous membrane, each perfused by its own channel. Lateral chambers apply cyclic mechanical strain.

How Does Organ-on-a-Chip Work?

A pump or pressure controller drives culture medium through the chip’s channels at a defined flow rate. Cells experience continuous nutrient delivery, waste removal, and shear stress. Where a flexible membrane is present, cyclic deformation can be added to reproduce mechanical strain such as breathing or peristalsis.

Flow rate is the primary control variable because it sets shear stress, and shear stress directly drives cell behaviour. Endothelial cells exposed to physiological shear elongate and align; the same cells under static conditions do not. Fluigent’s shear stress calculator converts channel geometry and flow rate into the resulting wall shear stress. Stability matters as much as magnitude. Pulsatile or drifting flow introduces variability that is difficult to distinguish from biological effects, which is why pressure-driven control is often chosen over peristaltic alternatives for long experiments. Our review of five key tips for starting organ-on-chip models covers the practical setup.

Figure 2: Perfusion circuit for an organ-on-a-chip experiment. Applied pressure drives medium through the chip at a flow-sensor-controlled rate; recirculating a defined reservoir volume allows secreted factors to accumulate.

History of Organ-on-a-Chip Technology

Organ-on-a-chip technology arose around 2010 from the convergence of semiconductor-derived microfabrication methods with mounting evidence of the limited translational predictivity of animal models. Building on lab-on-a-chip architectures, the field diverged by introducing living tissue interfaces and physiologically relevant mechanical stimulation — neither required in chemical microdevices.

Two pressures converged to make that step worth taking. Pharmaceutical attrition rates had stayed stubbornly high despite decades of methodological refinement, and the cost of discovering a human-specific toxicity in clinical trials rather than in the laboratory had become difficult to justify. At the same time, soft lithography had made microfluidic device fabrication cheap enough for a biology laboratory to iterate on a design in days rather than months.

timeline Development of organ on a chip technology

Figure 3: Development of organ-on-a-chip technology, from lab-on-a-chip microfabrication through the first breathing lung model to linked multi-organ systems and regulatory recognition.

From Lab-on-a-Chip to the First Breathing Lung

Microfluidic engineering long predates organ chips. Through the 1990s and 2000s, lab-on-a-chip devices miniaturised chemical and analytical operations — mixing, separation, detection — onto glass and polymer substrates. The manufacturing toolkit was mature well before anyone used it to culture tissue.

The turning point came in June 2010, when Huh and colleagues at Harvard’s Wyss Institute published a lung model in Science. Their device placed alveolar epithelial cells and pulmonary microvascular endothelial cells on either side of a thin porous membrane, then applied cyclic suction to side chambers so the membrane stretched and relaxed like breathing tissue [1]. The result was more than a structural mimic. The model reproduced organ-level responses to bacteria and inflammatory cytokines, and revealed that cyclic mechanical strain amplified the toxic and inflammatory response to silica nanoparticles — an effect confirmed in whole mouse lung. Mechanical motion was shown to change the biological response, which established stimulation as essential rather than decorative.

Toward Multi-Organ and Body-on-a-Chip Models

Single-organ devices multiplied quickly after 2010. Gut, liver, kidney, skin, bone marrow, heart, and blood-brain barrier models followed within a few years, each adapting channel design, matrix, and cell composition to a different physiological question. Sustained funding from national research agencies accelerated this expansion considerably, and by the end of the decade most major organs had at least one published chip model.

Attention then turned to connecting the different organ-on-a-chip models. Many pharmacological questions are inherently multi-organ. Intestinal absorption followed by hepatic biotransformation can generate a metabolite that, rather than the parent molecule, produces renal injury. Single-organ devices cannot reproduce this sequence, since the compound presented to the tissue is operator-defined rather than the product of upstream metabolic processing.

Linked models address this by circulating a shared medium between two or more chips. A liver–kidney organ-on-chip model built with the Omi Dual platform was developed with researchers at the University of Limoges to study how both organs jointly process tacrolimus. The ambition of a complete body-on-a-chip, with every major organ scaled and connected, remains a long-term research goal. Scaling ratios, a universal medium serving different tissues, and analytical access to very small volumes are all unresolved.

Organ-on-a-Chip vs Conventional Models

Organ chips occupy an intermediate position between two established approaches. They provide greater physiological fidelity than conventional static cell culture and greater experimental control than animal studies, while not reproducing the systemic completeness of an intact organism. They are therefore best characterized as a complementary methodology, addressing questions that the other two models resolve poorly

Compared With 2D Cell Culture

Conventional culture grows cells on flat plastic in static medium. It is inexpensive, well characterised, and scales to high-throughput screening, which is why it remains the default for early compound triage. Its limits are well documented: cells flatten into unnatural morphologies, lose tissue-specific functions, and never experience flow.

Hepatocytes are the standard illustration. In static culture, they progressively lose cytochrome P450 activity within days, so any metabolism study runs against a clock and long-term dosing is simply out of reach. Under continuous perfusion, the same cells maintain function for considerably longer, making chronic exposure studies feasible and allowing repeat dosing that mirrors how a patient would take a drug. There is a second, subtler difference. In a static well, a cell sits in its own accumulated secretions and a depleting pool of nutrients until the next medium change resets everything abruptly. The biochemical environment therefore swings between two extremes on a fixed schedule. Perfusion replaces that sawtooth with a steady state, which matters whenever the readout is sensitive to concentration — as autocrine and paracrine signalling almost always is.

The trade-off is throughput and cost. A well plate runs hundreds of conditions in parallel; a chip experiment runs a handful, requires fluidic connections, and demands more technical skill from the operator. Chips are best deployed where a mechanistic answer justifies the effort, not as a wholesale replacement for plate-based screening, and most laboratories run both.

Compared With Organoids and Spheroids

Organoids are self-organising 3D structures grown from stem cells that reproduce tissue architecture and cellular diversity with remarkable fidelity. Spheroids are simpler aggregates. Both deliver genuine three-dimensional biology that flat culture cannot.

Their common weakness is transport. As an organoid grows past a few hundred micrometres, diffusion alone no longer supplies its core, producing a necrotic centre that confounds any measurement averaged over the whole structure. Neither format has vasculature, and neither experiences directional flow — so neither can model drug delivery through a blood vessel, the very route by which most compounds reach tissue.

Reproducibility is a further difference. Organoids self-organise, which is their strength biologically and their difficulty experimentally: size, shape, and composition vary between individual structures within the same culture. A chip imposes geometry, so the architecture is defined by the device rather than emerging differently each time. The two approaches are increasingly combined rather than compared. Placing organoids inside a perfused chip supplies the missing transport and mechanical cues while keeping the self-organised architecture, and this hybrid is now among the more active directions in the field.

Comparison of three in vitro formats of nam

Figure 4: Comparison of three in vitro formats: static 2D monolayer culture, a self-organised organoid with a diffusion-limited core, and a perfused organ-on-a-chip with defined geometry and continuous flow.

Benefits of Organ-on-a-Chip Technology

The core benefit is human relevance under controlled conditions. Chips use human cells in a physiologically realistic environment while letting researchers set flow, geometry, and biochemical exposure precisely — a combination neither static culture nor animal studies provide. Three distinct advantages follow from this.

Human biology, tested directly. Species differences explain many late-stage failures: a compound safe in rodents can be toxic in humans because the relevant receptor, transporter, or metabolic pathway differs. Roughly nine in ten drug candidates that appear safe and effective in animals ultimately fail to reach approval, largely on human safety or efficacy grounds. Working with human cells from the outset removes the translation step.

Mechanical and biochemical control. Flow rate, shear stress, concentration gradients, and cyclic strain are all set by the operator and recorded. That control makes it possible to isolate one variable and attribute an observed change to it, which is difficult in an animal where systemic compensation obscures cause and effect.

Access to the experiment. Chips are optically transparent and thin enough for high-magnification objectives, so live imaging runs throughout the experiment rather than only at endpoint. Medium can be sampled repeatedly for secreted factors, giving a full time course from a single device instead of a series of sacrificed replicates. This changes the statistics as well as the convenience: each chip becomes its own baseline, and the variability between biological replicates no longer masks the effect being measured.

Small volumes, concentrated signals. Because the circulating volume is measured in microlitres to a few millilitres, a secreted biomarker reaches a detectable concentration far sooner than it would in the dilution of a culture flask. Compound consumption falls in proportion, which matters early in development when only milligrams of a candidate molecule exist.

Patient-specific models. Because induced pluripotent stem cells can be derived from an individual donor, a chip can carry that person’s genotype. This supports the study of rare disease phenotypes and of variable drug responses across a population, and points toward genuinely personalised medicine.

Ethical and regulatory alignment. The regulatory environment has shifted decisively. The FDA Modernization Act 2.0, signed in December 2022, amended the Federal Food, Drug, and Cosmetic Act to replace the requirement for animal testing with “nonclinical tests,” explicitly naming cell-based assays, organ chips, and microphysiological systems as qualifying alternatives [3,4]. Our review of organ-on-chip in drug development examines what this means in practice.

Limitations of Organ-on-a-Chip Systems

The technology is not a finished replacement for existing models. Material chemistry can distort drug studies, standardisation across laboratories is incomplete, throughput remains low, and no chip reproduces the immune, endocrine, and neural integration of a living body. These constraints are actively worked on and worth understanding before committing to the approach.

Material Absorption and Chip Chemistry

Polydimethylsiloxane (PDMS), the elastomer used in the majority of academic devices, is optically transparent, gas-permeable, and readily replica-molded, which accounts for its prevalence. It also absorbs small hydrophobic molecules into its bulk, a class that includes many drug candidates.

The consequence is a measurement error in the direction that matters most. If a fraction of the dose partitions into the chip walls, the concentration reaching the cells is lower than the nominal dose, and potency is systematically underestimated. Absorption in PDMS devices is a recognised limitation in the microphysiological literature [5].

Mitigations exist: surface coatings that block partitioning, alternative thermoplastics such as polycarbonate or cyclic olefin copolymer, and mass-manufactured devices designed from the outset for pharmacological work. Each involves a trade. Thermoplastics are not gas-permeable in the same way, so oxygenation must be engineered elsewhere in the circuit, and they are harder to prototype in a university workshop than a moulded elastomer. The practical advice is to characterise the chip before trusting a dose-response curve from it. Running the compound through the device without cells and measuring what emerges at the outlet establishes how much is lost to the walls, and that correction can be applied to everything measured afterwards.

Standardisation, Throughput and Cost

Reproducibility across laboratories remains the field’s most cited obstacle. Chip geometries, coating protocols, cell sources, and flow regimes vary between groups, so comparing two published models is rarely straightforward and inter-laboratory validation is demanding.

Throughput is a second constraint. Every chip needs connection, priming, bubble management, and monitoring. A skilled operator runs far fewer conditions than a plate-based assay allows, which restricts chips to focused mechanistic questions rather than broad screening campaigns. Cost follows from both. Instrumentation, single-use consumables, and specialised training carry real expense, and the analytical side adds more: sampling microlitre volumes and detecting biomarkers at those scales requires sensitive methods and careful handling. Automation and standardised consumables are steadily reducing this burden, but a chip programme still represents a genuine capital and training commitment rather than an incremental purchase.

Finally, no chip reproduces a whole body. Systemic immune surveillance, hormonal feedback, neural control, and whole-body pharmacokinetics emerge from an integrated organism and cannot be recovered by connecting a few compartments. Any endpoint that depends on those interactions still requires an animal model or a clinical study, and honest scoping of what a chip can and cannot answer is the single best predictor of whether a project succeeds.

Common Applications of Organ-on-a-Chip Systems

Organ-on-a-chips are used wherever human-specific biology under flow determines the answer: drug development and toxicity testing, disease modelling, barrier and transport studies, and personalised medicine. The most established use is preclinical safety assessment, where predicting human hepatotoxicity or cardiotoxicity before clinical trials has clear economic and ethical value.

Drug development and toxicity. Liver chips assess drug-induced injury, cardiac models detect arrhythmia risk, and kidney models examine transporter-mediated clearance. Our drug development review covers this area in depth.

Disease modelling. Tumour-on-chip models reproduce the microenvironment and vascular delivery that shape treatment response; see our cancer-on-chip review. Infection, inflammation, and fibrosis models follow similar logic.

Barrier and transport studies. Gut, blood-brain barrier, and lung interfaces are natural chip applications because their function is defined by a two-compartment interface under flow. A breathing lung-on-chip case study illustrates the approach.

Omi: Fluigent’s Organ-on-a-Chip Platform

Omi is Fluigent’s automated perfusion platform for organ-on-a-chip work. It controls the fluidic environment around cells inside a microfluidic chip, handling perfusion, recirculation, sampling, and injection in one device, with flow rate control from 1 µL/min to 1 mL/min and a maximum pressure of 600 mbar.

Recirculation is central to long-term culture. A 4 mL reservoir volume can be circulated unidirectionally for days or weeks at a defined flow rate, so secreted factors accumulate as they would in a closed physiological compartment instead of being diluted at every medium change. The platform is designed for the realities of a cell biology laboratory rather than an optical bench. It fits inside an incubator, several units share one incubator, and a two-hour battery lets a running experiment be carried to a microscope for live imaging without interrupting perfusion. Fresh medium, cells, or test compounds are injected and samples collected without disconnecting the chip, which limits contamination risk.

Two configurations cover most models. Single mode perfuses one channel — the format used for endothelial and vascular work. Dual mode synchronises two units on one chip to serve both compartments independently, which is what a blood-brain barrier or any liquid–liquid interface model requires. Omi is compatible with essentially any chip design, and the Organ on Chip Perfusion Pack bundles the components needed for a first perfused culture.

Download the Free Organ-on-a-Chip Whitepaper

Fluigent’s white paper, A Review of Organ on Chip Technology, sets out the field in more depth than a web guide allows: device architectures, organ models, flow control considerations, and the published work behind them. Download the full white paper — or request a demo or quote for Omi if you would rather discuss a specific model with an application scientist.

Frequently Asked Questions

What Equipment Is Required for Organ-on-a-Chip Experiments?

A minimum setup needs four things: a microfluidic chip suited to the model, a flow control instrument, tubing and connectors compatible with sterile culture, and an incubator. Flow control is the component that most affects data quality, because unstable flow produces variability indistinguishable from biological effect. Laboratories generally take one of two routes. An integrated platform such as Omi combines pressure control, sensing, and protocol automation for perfusion and recirculation in a single incubator-compatible unit. Alternatively, a modular setup built around a Flow EZ pressure controller offers more flexibility for custom fluidic circuits. Beyond that, expect an inverted microscope for imaging and standard sterile handling equipment.

What Is Organ-on-a-Chip Toxicity Testing?

Organ-on-a-chip toxicity testing exposes human cells cultured under perfusion to a compound and measures the damage. Because the cells are human and the environment is physiological, the result predicts clinical toxicity more directly than an animal study or a static assay does. Liver chips are the most developed case, since drug-induced liver injury is a leading cause of late-stage drug attrition and of post-approval withdrawal. Cardiac models assess arrhythmia risk and kidney models examine transporter-mediated injury. Practically, testing means dosing the perfused chip over hours to weeks while tracking viability, released biomarkers such as transaminases, and functional readouts such as barrier integrity. The repeated-sampling capability of a recirculating platform makes the time course accessible from one device.

What Is the Difference Between Organ-on-a-Chip and Lab-on-a-Chip?

Lab-on-a-chip is the broader category: any microfluidic device that ptimizedd laboratory operations, including mixing, separation, PCR, and analytical detection. Organ-on-a-chip is a ptimizedd subset in which the device cultures living cells to reproduce the function of a tissue or organ. The practical distinction is what the device is ptimized for. A lab-on-a-chip handles fluids and samples, and its design constraints are chemical and analytical. An organ chip must additionally sustain living tissue, which brings requirements a chemical device never faces: biocompatible materials, sterility over weeks, oxygen supply, and physiologically appropriate shear stress. Every organ-on-a-chip is a lab-on-a-chip; the reverse is not true.

Can Organ-on-a-Chip Technology Replace Animal Testing?

Not entirely, and not yet — but the regulatory direction is clear. The FDA Modernization Act 2.0 removed the statutory requirement for animal testing in investigational new drug applications and named organ chips as a qualifying alternative [3]. In April 2025 the agency published a roadmap aiming to make animal studies the exception within three to five years, and the Senate passed a follow-up bill in December 2025 [6]. Legal permission is not scientific readiness. Chips do not reproduce systemic immune responses, whole-body pharmacokinetics, or behavioural endpoints, and validation packages proving equivalence for specific contexts of use are still being assembled. The realistic near-term outcome is substantial reduction and targeted replacement in areas such as hepatotoxicity, rather than wholesale substitution. Our drug development review follows this question in detail.

Conclusion

Organ-on-a-chip technology reproduces the function of human tissue by combining living cells, a microfluidic device, and continuous perfusion. It answers questions that flat culture and animal models answer poorly — human-specific toxicity, barrier transport, and mechanically driven biology — without replacing either for systemic endpoints.

The realistic position today is complementary rather than substitutive. Regulatory frameworks now recognise microphysiological data explicitly, and validation is accumulating fastest where the failure mode is human-specific, such as drug-induced liver injury. Material chemistry, standardisation, and throughput remain genuine constraints, and a project succeeds or fails largely on whether its scope matches what a chip can actually measure.

For a laboratory starting out, the practical entry point is one well-characterised model run under stable, documented perfusion rather than an ambitious multi-organ design. Our five key tips for organ-on-chip models covers the first setup, and the Organ on Chip Perfusion Pack bundles the components needed to run it.

Related Expertises

References

Huh D, Matthews BD, Mammoto A, Montoya-Zavala M, Hsin HY, Ingber DE. Reconstituting organ-level lung functions on a chip. Science. 2010;328(5986):1662–1668. doi:10.1126/science.1188302

Dzakah EE, Ni C, Feng X, Hu W, Zhao B, Tang S, Pang C. Modernizing drug development and ensuring global equity in the organoid revolution. Commun Biol. 2026 Apr 17;9(1):540. doi: 10.1038/s42003-026-10076-4.

FDA Modernization Act 2.0: transitioning beyond animal models with human cells, organoids, and AI/ML-based approaches. Journal of Clinical Investigation. 2023. PMC10617761 [TO VERIFY: full author list]

Han JJ. FDA Modernization Act 2.0 allows for alternatives to animal testing. Artif Organs. 2023 Mar;47(3):449-450. doi: 10.1111/aor.14503.

van Meer BJ, de Vries H, Firth KSA, van Weerd J, Tertoolen LGJ, Karperien HBJ, Jonkheijm P, Denning C, IJzerman AP, Mummery CL. Small molecule absorption by PDMS in the context of drug response bioassays. Biochem Biophys Res Commun. 2017 Jan 8;482(2):323-328. doi: 10.1016/j.bbrc.2016.11.062.

US Food and Drug Administration. Roadmap to Reducing Animal Testing in Preclinical Safety Studies. April 2025.

What Is Organ-on-a-Chip? A Complete Guide