Microfluidic devices, or lab-on-a-chip technology, typically involve minimal fluid control [1]. These microfluidic devices measure and analyze the analyte interactions in fluid samples, such as biomolecules, cells, tissues, or detection devices (Fig. 1). Additionally, they can be used to synthesize compounds by taking advantage of interactions that occur in microfluidic systems.

Figure 1. Microfluidic device applications in medicine.

In biology, drug research often relies on cell-based analysis systems. Static cell assays primarily use one type of cell or tissue, making it impossible to evaluate various physiological interactions that occur due to the exchange of metabolites between cells and tissues. While animal experiments can supplement these limitations, they are expensive and raise ethical concerns [2]. Thus, recent research has actively pursued the development of biomimetic systems based on microfluidics to overcome the challenges mentioned above. These systems offer advantages such as reduced sample and reagent usage, shorter analysis times, and results that closely mimic those obtained from in vivo experiments.

Microfluidic technology is not only used for analysis but also for drug synthesis [3]. Unlike conventional fluid dynamics, laminar flow with uniform characteristics, free from turbulence, can be achieved in microfluidics. The easy control of fluid flow properties allows for stable concentration gradients, droplet formation, mixing of two or more fluids at uniform ratios, and the operation of mixers, among other applications. This review focuses on how microfluidic devices have been utilized in drug research and biomimetic modeling for in vitro studies over the past five years.

The authors searched the PubMed database for papers from the last 5 years using the keywords: microfluidic, in vitro, biomimetic, drug, therapy, and not review. As a result, 26 papers were sourced, of which 19 were reviewed. Systematic review and meta-analysis were excluded. The authors selected and prepared papers related to this content, including research topics, titles, and methods.

Microfluidic device fabrication methods

Among the 19 papers reviewed (Table 1), PDMS (Polydimethylsiloxane) was used in 10 papers [4-13], making it the most used material for microfluidic devices. PMMA (poly(methyl methacrylate)) and polycarbonate materials were each used in 2 papers [6,7,9,14], and following them, various other materials such as PEEK (polyetheretherketone) [15], UV resin [15], PET (polyethylene terephthalate) [9], GelMa (gelatin methacryloyl) [16], ECM gel [5], agarose [17] and polystyrene [18] were used in one paper each. Additionally, there were 2 papers in which the device material information was not specified [19,20]. Among the 19 papers, 5 described using devices made from two or more composite materials [5-7,9,15].

Table 1 . Selected and reviewed articles on microfluidics from PubMed

Title	Cell/tissue type	Material	Structure of device	Objective of device	Result	Reference
Colorectal tumor-on-a-chip system: a 3D tool for precision onco-nanomedicine	Human colorectal tumor cells (HCT-116), human colonic microvascular endothelial cells (HCoMEC)	PDMS	One circular chamber, two lateral channels	Emulating the human colorectal tumor microenvironment and enabling the reconstitution of physiological functions of microvascular tissue	They developed a 3D microfluidic chip that mimics the human colorectal tumor microenvironment and allows the generation of a gradient of drug-loaded NPs 1	[4]
Patient-specific organotypic blood vessels as an in vitro model for anti-angiogenic drug response testing in renal cell carcinoma	Primary endothelial cells from patient, HUVECs, primary monocytes	PDMS, ECM gel	Central channel through the hydrogel	To create an organotypic in vitro vessel model with primary patient-specific endothelial cells that can mimic the structure and function of normal and tumor-associated blood vessels	Their microfluidic model recapitulated the functional and structural differences between TEnCs and NEnCs, and provided a platform for personalized drug screening and assessment	[5]
Near-physiological microenvironment simulation on chip to evaluate drug resistance of different loci in tumour mass	HepG2, human breast cancer cell line (MCF-7), MCF-7/ADR, HUVECs, mouse fibroblast cell line (NIH-3T3)	PDMS, polycarbonate	Sandwich structure of double channel, porous membrane and wide single channel	To simulate the tumour-vascular microenvironment and evaluate the drug resistance of different loci in the tumour mass	They successfully built a tumour-vascular microenvironment consisting of four types of cells: endothelial cells, fibroblasts, breast tumour cells and their resistant branch	[6]
Microengineered human blood-brain barrier platform for understanding nanoparticle transport mechanisms	Human brain microvascular endothelial cells (HBMECs), human brain pericytes, human astrocytes	PDMS, polycarbonate	Sandwich structure of single channel, porous membrane and double channel	To create a microfluidic model of the human BBB that can mimic the structure and function of the brain microvasculature and enable 3D mapping of nanoparticle distributions in the vascular and perivascular regions	They created a human BBB model that mimics the key structure and function of the BBB, such as cellular interactions, gene expressions, low permeability, and 3D astrocytic network	[7]
Engineered biomimetic nanoparticle for dual targeting of the cancer stem-like cell population in sonic hedgehog medulloblastoma	Medulloblastoma cells, SmoA1-GFP, PTC knockout cells, DAOY, PZp53	PDMS	3 Inlets, mixer	To synthesize eHNPs with high homogeneity and reproducibility	They were able to synthesize eHNPs with high homogeneity and reproducibility in a single-step process	[8]
Metastasis-on-a-chip mimicking the progression of kidney cancer in the liver for predicting treatment efficacy	HepLL, Caki-1	PMMA, PDMS, PET	Several holes linked to central circular chamber	To create a 3D metastasis-on-a-chip model that can mimic the progression of kidney cancer cells in the liver and predict the treatment efficacy of anti-cancer drugs	The authors successfully created a 3D metastasis-on-a-chip model with organ-specific ECM by co-culturing kidney cancer cells (Caki-1) and hepatocytes (HepLL) in a decellularized liver matrix (DLM)/GelMA-based hydrogel	[9]
Evaluation of hepatic drug-metabolism for glioblastoma using liver-brain chip	Glioblastoma cells (U87), HepG2, brain microvascular endothelial cells (BMECs)	PDMS	Sandwich structure of single channel and double channel	To create a liver-brain chip model that can simulate the physiological and pharmacological processes of oral anti-brain-tumor drugs in vitro	They could simulate the process of oral anti-brain-tumor drugs delivery in vitro	[10]
Parallel and large-scale antitumor investigation using stable chemical gradient and heterotypic three-dimensional tumor coculture in a multi-layered microfluidic device	Human glioma cells (U251), mouse embryonic fibroblasts (NIH 3T3)	PDMS	Gradient maker, arrays of chambers	To create a microfluidic platform for drug screening in a 3D cancer microenvironment	They successfully produced array-like and size-homogeneous heterotypic 3D tumors composed of glioma cells and fibroblasts, which showed in vivo-like features such as tumor-stromal interaction, metabolic activity gradient, and viability gradient	[11]
Microfluidic mixing system for precise PLGA-PEG nanoparticles size control	Endothelial cells (EA.hy926), macrophage (PMA-differentiated THP-1)	PDMS	Mixer with micro channel	To produce polymeric NPs with well-controlled sizes and high monodispersity	The authors were able to produce polymeric NPs with well-controlled sizes ranging from 30 to 70 nm by varying the polymer concentration, flow rate and flow rate ratio between the aqueous and organic solutions	[12]
Microfluidic arrays of breast tumor spheroids for drug screening and personalized cancer therapies	Breast cancer cells (MCF-7), patient-derived breast cancer cells	PDMS	Arrays of chambers, gradient maker	To create a microfluidic platform for the generation of large arrays of breast tumor spheroids that are grown under close-to-physiological flow in a biomimetic hydrogel	They showed that the platform can support the growth of patient-derived spheroids from different breast cancer subtypes and correlate their drug response with in vivo data	[13]
Microfluidic fabrication of ph-responsive nanoparticles for encapsulation and colon-target release of fucoxanthin	Human colon adenocarcinoma cells (Caco-2)	PMMA	Mixer	To synthesize NPs of fucoxanthin, a marine natural pigment with various biological activities	The authors successfully synthesized biocompatible fucoxanthin in shellac NPs (FX/SH NPs) with a uniform size and a high loading capacity for colon-targeted delivery of fucoxanthin	[14]
A 3D biohybrid real-scale model of the brain cancer microenvironment for advanced in vitro testing	Human cerebral microvascular endothelial cell line (hCMEC/D3), primary human astrocytes, glioblastoma cells (U87)	PEEK, UV curable resin (AA 3494 Loctite)	Parallel porous capillaries	To present a 3D biohybrid model of the brain cancer microenvironment	They proved the feasibility of a triple co-culture of endothelial cells, astrocytes, and magnetically-driven spheroids of glioblastoma cells inside the microfluidic device, forming a biohybrid BBB	[15]
Ball-bearing-inspired polyampholyte-modified microspheres as bio-lubricants attenuate osteoarthritis	Chondrodyte cells	GelMA	Sphere porous structure	Enhancing joint lubricartion, anti-inflammatory therapy	They grafted poly(dopamine methacrylamide-to-sulfobetaine methacrylate) (pSBMA) brushes onto MGSs to form superlubricated MGSs (MGS@DMA-SBMA), which could further reduce the friction coefficient due to the hydration lubrication mechanism	[16]
Targeting magnetic nanoparticles in physiologically mimicking tissue microenvironment	Human colorectal tumor cells (HCT 116)	Agarose	Central channel through the hydrogel	To demonstrate a microfluidic platform that can mimic the human vasculature and the surrounding tissue, and study the transport and penetration of magnetic nanoparticles (MNPs) across the deformable	They fabricated hydrogel-based microchannels that mimicked the human vasculature and tissue stiffness, and studied the magnetically assisted penetration of NPs across the deformable microenvironment under physiological flow rates	[17]
An integrated biomimetic array chip for establishment of collagen-based 3D primary human hepatocyte model for prediction of clinical drug-induced liver injury	PHHs	Polystyrene	2 Concentric superimposed holes	To develop 3D PHH model for the prediction of clinical drug-induced liver injury	They optimized the collagen-based 3D PHH model on the iBAC and showed that it had improved and stabilized liver functionality, such as cell viability, albumin production, urea secretion, and metabolic enzyme activity	[18]
ALTEN: a high-fidelity primary tissue-engineering platform to assess cellular responses ex vivo	Mammary adenocarcinoma cell lines (4t1.2, 67NR), primary mammary tumor biopsies, human gastric tumor biopsies, human breast tumor biopsies	Not presented in paper	3 Inlets, strainer and droplet generator	Encaptulating tissue fragments of different sizes with alginate hydrogel	They were able to produce shape-controlled hydrogels of different sizes and geometries that encapsulated tissue fragments from various organs and species	[19]
Microfluidic sonication to assemble exosome membrane-coated nanoparticles for immune evasion-mediated targeting	Human lung carcinoma cells (A549), mouse macrophage-like cells (RAW264.7)	Not presented in paper	3 Inlets, double spiral channels	To synthesize biomimetic NPs with natural membrane coating for enhanced tumor targeting and drug delivery	They successfully fabricated EM-PLGA NPs, CCM-PLGA NPs, and lipid-PLGA NPs with similar sizes and core−shell structures in a one-step and continuous manner	[20]
Efficacy of molecular and nano-therapies on brain tumor models in microfluidic devices	Human glioblastoma multiforme (U87-MG) cell lines, primary human astrocytes	PDMS	Single channel device and double channel device	To use microfluidic devices to model the in vitro spatial organization of brain tumors and to test the anti-cancer efficacy of different therapeutic agents, such as free docetaxel, docetaxel-loaded NPs, and Fmoc-Glc6P	In the single-channel device, DTXL had a higher cytotoxic effect on 3D organized U87-MG cells than on 2D cell monolayers, with 50-fold lower IC50 values. In the double-channel device, DTXL and DTXL-SPN had comparable but lower cytotoxicity than in the single-channel device, suggesting that the vascular/tissue interface reduced the drug delivery to the tumor cells	[37]
Function of hepatocyte spheroids in bioactive microcapsules is enhanced by endogenous and exogenous hepatocyte growth factor	Hepatocytes	PDMS	Droplet generator	To enhance the phenotype and function of hepatocyte spheroids	They successfully fabricated bioactive core-shell microcapsules with heparin-containing hydrogel shells and aqueous cores using a co-axial flow-focusing microfluidic device 1	[38]

3D, three-dimensional; PDMS, polydimethylsiloxane; NPs, nanoparticles; HUVECs, human umbilical vein endothelial cells; HepG2, human liver cancer cell line; BBB, blood-brain barrier; eHNPs, engineered high-density lipoprotein-mimetic nanoparticles; PMMA, poly(methyl methacrylate); PET, polyethylene terephthalate; GelMA, gelatin methacryloyl; PEEK, polyetheretherketone; MGSs, microfluidic gelatin methacrylate spheres; PHH, primary human hepatocyte; iBAC, integrated biomimetic array chip.

Most of the devices were fabricated by soft lithography (Fig. 2). Soft lithography is the most widely used method for microfluidic fabrication. It involves creating molds on a silicon wafer using a photoresist to pattern the surface, then pouring and curing a polymer like PDMS on top of the mold. After the curing process is complete, the polymer is peeled off, and the device is completed by bonding it to glass. During bonding, glue can be used, or the surfaces can be treated with plasma for adhesion [21].

Figure 2. Schematic of soft lithography protocol used in the fabrication of microfluidic devices.

Microfluidic device designs

Just as diverse as the materials and applications of microfluidics are, the designs in microfluidics vary greatly. However, some commonly used designs include droplet generators, mixers, gradient makers, and sandwich types [22,23]. Droplet generators typically utilize a T-junction [24] configuration and employ two or more distinct, non-mixing fluids to create droplets of the desired size (Fig. 3).

Figure 3. T-junction droplet generators. (A) Regular T-junction geometry where the dispersed phase is injected perpendicularly into a continuous fluid stream. (B) A frequently used device in which the continuous phase narrows the filament of the dispersed liquid down until pinch-off takes place.

Mixers are the simplest yet powerful devices in microfluidics [25,26]. They can create uniformly mixed substances solely through diffusion occurring in microchannels (Fig. 4). On the other hand, a gradient maker utilizes a complex structure, incorporating the principles of mixers to generate fluid concentration gradients or synthesize two or more fluids with gradient concentration levels [1,25,27]. These systems are often combined with an array of chambers, allowing simultaneous screening of various concentrations across multiple samples, thereby enabling the acquisition of a substantial amount of data using a minimal sample [28-32].

Figure 4. (A) Most commonly used designs of mixer. (B) Gradient maker design that is made of combinations of mixers.

Sandwich types using microgrooves or porous layers are also frequently employed in microfluidics [33-36]. These designs typically consist of multiple channels stacked in the horizontal or vertical direction, with porous membranes or microgrooves acting as barriers between each channel, serving a filtering function. Different cells or solutions can be introduced into each channel, and the membranes or grooves between the channels allow for selective passage of substances of the desired size [2].

Microfluidic systems for cancer studies

Among the 19 reviewed papers, 14 of them were focused on tumor or cancer-related research. Among these, brain tumors and cancers were the most studied topics, with 5 papers dedicated to them [7,8,10,11,15]. Colorectal cancer-related research followed with 3 papers [4,14,17], while breast cancer [6,13] and kidney cancer [5,9] each had 2 papers. Only 1 paper was tributing lung cancer [20]. There was 1 paper targeting various types of cancers [19]. Of the 6 brain-related studies, 5 commonly addressed the blood-brain barrier (BBB) [7,8,10,15,37], while the remaining 1 study focused on the spheroid model [11].

Biomimetic microfluidic devices for drug research

Of the 19 reviewed studies, 14 focused on drug screening and applied microfluidic devices as a modeling strategy. Among these, 10 studies were specifically targeting drug screening for cancer-related drugs [4-13]. The remaining 2 studies were focused on research involving only hepatocytes [18,38]. In the perspective of drug synthesis, 4 of 19 papers were tributing drug synthesis. Three studies were synthesized nanoparticles [9,13,15], and 1 made joint lubricant for osteoarthritis therapy [16].

Microfluidics are custom-made based on the experimental field and purpose, leading to a wide variety of designs and applications. Hence, we explored applications concerning drug research and biomimetic, which are topics often cited and recognized in the context of microfluidics. Our investigation emphasized materials, cell types, and design aspects, utilizing PubMed as the primary platform for our exploration.

From a materials perspective, most of the reviewed studies primarily used PDMS. This choice is likely due to PDMS’s gas permeability, transparency, and ease of molding, making it suitable for cell culture and observation. Other materials used in these studies exhibited similar properties to PDMS [39].

Most of the published articles reviewed were related to cancer research, particularly in the context of cancer or tumor vasculature. In brain-related research, there was a significant emphasis on studies related to the blood-brain barrier [7,8,10,15,37]. The BBB is a specialized structure in brain tissue directly related to drug delivery and brain injury [40]. To pass through the BBB, small particles like nanoparticles are advantageous. This issue explains the numerous attempts to create experimental models resembling the vascular barrier for nanoparticle screening.

Microfluidics also can be a useful tool for spheroid studies [11,13,19,38]. These studies often involved the cultivation of multiple spheroids in an array format or the use of gradient makers to screen drugs at different concentrations simultaneously. Microfluidics has demonstrated high reproducibility in biomimetic modeling, including cancer models. Adopting a sandwich structure enables the creation of three-dimensional models, moving beyond the limitations of flat, two-dimensional models [6,7,10]. Using porous structures [6,7,9,15,16] or microscale channels [4], microfluidics can accurately mimic the transport of metabolites, closely resembling the processes that occur within the human body. Microfluidic systems not only function as a structural framework but also as a versatile tool for model creation. It presented a consistent re-size process of spheroid or tissue [19].

Regarding drug synthesis, research employing droplet generators has exhibited exceptional stability and performance in size control and encapsulation. They offer a favorable environment for synthesizing small particles, such as nanoparticles. The primary device designs included droplet generator designs [19,38] and mixer designs [8,12,14].

Droplet generators are specialized microfluidic devices designed to control small-sized particles or droplets of fluid. These generators create droplets by utilizing immiscible fluids flowing from 2 or 3 inlets and a junction. By adjusting the velocity, phase, and channel width of the fluids entering each inlet, along with the width of the outlet pinch, droplet generators can produce droplets of desired sizes or encapsulate drugs.

Mixer designs, although simple in structure, have proven highly effective in simultaneously diluting two or more drugs. Microfluidic mixers typically feature multiple inlets and a single outlet. Depending on experimental requirements, they may incorporate various curves within the channels or grooves on the device’s bottom surface to enhance mixing. The key distinction between microfluidic mixers and traditional mixing methods lies in the fact that, in a microfluidic environment, the diffusion of molecules allows for uniform and consistent mixing over a smaller area. Consequently, microfluidic mixers offer significant advantages over traditional mixing methods by reducing the time and equipment required for sample mixing. This is particularly advantageous when dealing with samples or substances with high costs or when changes in properties due to heat need to be minimized.

Overall, the advantages of employing microfluidics as a platform for biomimetic studies, particularly in drug development and cancer research, were highly evident. The development and application of microfluidic systems in other research and medical fields will continue in the following decades.

Microfluidic devices have demonstrated remarkable performance in biomimetic modeling and drug research, with notable strengths in cancer modeling, parallel screening, and nanoparticle synthesis. Microfluidic platforms have consistently shown high reproducibility. However, while microfluidic technology has made significant advancements, there is still ample room for improvement. Areas that require further attention include expanding the diversity of materials, enhancing throughput capabilities, and reducing costs. Continuous development of microfluidic platforms is essential to meet the evolving demands of various biological studies.

None.

None.

Conceptualization: SHW. Data curation: HHS. Formal analysis: HHS. Funding acquisition: SHW. Investigation: HHS. Methodology: HHS. Software: HHS. Validation: SHW. Visualization: HHS. Writing–original draft: HHS. Writing–review & editing: all authors.

No potential conflict of interest relevant to this article was reported.

This work was supported by the Dankook Institute of Medicine & Optics in 2023. This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (RS-2023-00247651 and NRF-2020R1A6A1A03043283), Korea Medical Device Development Fund grant funded by the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, the Ministry of Food and Drug Safety (RS-2020-KD000027), Republic of Korea.

None.