2.1 Bioprinting

The conventional fabrication techniques of OOCs include photolithography, soft lithography,^[89-91] replica molding,^[92] capillary molding,^[89] microcontact printing,^[92] microtransfer molding,^[89] and injection molding.^[93-95] One of the limitations of the above methods is the limited fabrication capability for forming the complex structures of organs and tissues.^{[96, 97]} On the other hand, these methods usually require multistep production protocols. In particular, lithographic techniques need to be carried out through several lithographic processes and masks.^{[29, 96, 98]} This leads to experiments that are time consuming and expensive. In addition, the traditional methods require a secondary organization for the cell seeding process, which oftentimes drives up the overall cost, as well as poor selectivity of different cell types.

The adoption of 3D printing and 3D bioprinting in medical and biomedical applications has resulted in cost efficiency, rapid turnaround times, and a wide range of materials.^[99-103] Moreover, prototyping of OOCs with 3D bioprinting requires minimal microfabrication skills and enables simultaneous/consecutive (bio)printing of polymers, hydrogels, and multiple cell types to produce customized, reproducible, perfusable, and complex patient-specific 3D biomimetic tissue constructs with high precision in the placement of cells,^{[8, 29, 43, 96]} which is hardly achievable with the conventional techniques. 3D bioprinting techniques can be primarily divided into two categories (Figure 2): (i) nozzle-based (e.g., inkjet-based (droplet-based) and extrusion-based bioprinting) and (ii) light-enabled bioprinting (e.g., stereolithography apparatus (SLA)-/digital light processing (DLP)-based bioprinting, two-photon polymerization (TPP)-based bioprinting, laser-assisted bioprinting, and computed axial lithography).^[104-106] A summary of the advantages, limitations, and important properties of the commonly used 3D bioprinting methods is shown in Table 1.

2.2 Bioink

Solidifiable materials, including ion-crosslinkable hydrogels, temperature-sensitive polymers, and photopolymer bioinks are commonly used materials for 3D bioprinting.^{[2, 158]} The bioink can be derived from natural sources, such as alginate (also termed algin or alginic acid), carrageenan (also termed carrageenin), gellan gum, agar (also termed agarose), collagen, fibrin, gelatin, silk, fibrinogen, chitosan, methyl cellulose (derived from cellulose), and hyaluronan,^[159-162] or it can be derived synthetically, such as poloxamer (also termed Pluronic), Matrigel, poly(caprolactone) (PCL), poly(ethylene glycol) (PEG), GelMA, poly(vinyl alcohol) (PVA), poly(lactic acid) (PLA), and poly(lactic-co-glycolic acid) (PLGA).^{[161, 162]}

Alginate (a natural polymer extracted from brown seaweeds) is a cost-efficient option for bioink preparation with no toxicity, and good printability, yet with low cellular adhesion and slow degradation.^[163] Although carrageenan (a natural polysaccharide extracted from red seaweeds) is another bioink with fine biocompatibility, bioactive, mechanical, and rheological properties, its rather high toxicity is challenging.^[164-166] Gellan gum (an anionic microbial polysaccharide) is a cost-efficient biomaterial that offers shear-thinning properties, and high gelling efficiency, while suffering from limited printing fidelity.^[167] Agar (a polysaccharide isolated from sea kelp) is another bioink candidate that, despite having high mechanical strength and cost-efficiency, has limited cell adhesion.^[168] Collagen (one of the well-known body proteins) also can be used as bioink, offering good cell adhesion as well as growth, and facing challenges of low viscosity and weak mechanical properties.^[161] However, although gelatin has the same pitfalls as collagen, gelatin as a collagen-degradation product, offers nonimmunogenicity while maintaining the cell-friendly binding regions.^[161] Silk (a natural protein fiber) possesses biodegradability, nontoxic degradation residues, and high cellular viability, whereas it suffers from limited cell growth and function.^{[61, 161, 169]} Matrigel (solubilized basement membrane matrix secreted by Engelbreth-Holm-Swarm mouse sarcoma cells), although being expensive and unsuitable for translation into clinical setups, promotes cell growth and differentiation.^[170] PVA is a biocompatible, water-soluble polymer with suitable hydrophilicity and toughness for bioink preparation that has limitations in the adhesion of cells.^{[161, 171]} PEG (one of the FDA-approved materials for medical applications) is a biocompatible and easily modifiable polymer, well known for usage as sacrificial bioink, with limitations in mechanical strength and cell adhesion.^{[161, 169]} Furthermore, GelMA is a biocompatible and biodegradable option for the preparation of bioinks, where a crosslinking process with UV light is needed that can adversely affect cell viability, although visible-light options are becoming more broadly adopted.^{[161, 172]}

One of the most important factors in selecting one of these materials for the preparation of the desired OOC is the crosslinking mechanism, which should be chosen to be harmless to the cells to be loaded into the bioink, based on the fabrication method (photosensitive bioinks for light-induced methods and bioinks with shear-thinning ability for nozzle-based approaches). Cell interaction is another decisive factor for material selection. For example, bioinks made of collagen and hyaluronic acid are superior to alginate and silk fibroin, as they provide better cell interactions.^[159] For cancer modeling on a chip, the GelMA hydrogel is a suitable candidate as it promotes cell functions (e.g., tumor cell metastasis and invasiveness), rapid crosslinking, and biocompatibility.^[17] Biodegradability, on the other hand, may be important for certain applications. In these cases, it is recommended to use materials with a higher degradation rate compared with other bioinks, such as hyaluronic acid. In addition, silk fibroin usually shows better mechanical properties compared with the other bioinks.^[159] Bioinks should be able to maintain their mechanical integrity over a long incubation period under culture conditions (e.g., pressure, temperature, and humidity).^[173] An important role of bioinks is to encapsulate the cells to protect them during the bioprinting process, particularly in nozzle-based processes. Hybrid bioinks are promising innovations to improve the mechanical properties of available bioinks.^{[174, 175]} For instance, pure alginate, although readily processable, biocompatible, and widely available, has poor pattern fidelity and printability. However, the combination of alginate with cellulose nanocrystals (whisker/rod-shaped nanoparticles extracted from the crystalline regions of cellulose fibers and characterized by renewability, high mechanical strength, low density, and cytotoxicity) results in a reinforced bioink with better cell protection as well as shear-thinning properties that can be bioprinted through a 100-μm nozzle without clogging.^[176] In addition, composite formulas incorporating biologically derived components and living cells, called “living materials”, have recently gained attention for the specific characteristics that they offer, such as autonomic iridescence capability,^[177] especially for the preparation of bioinks for 3D bioprinting processes.^[178] Powered by various branches of sciences, such as microfluidics, genetics, and cell coating, living materials enable several applications in living structures and organ models.^{[178, 179]}

2.3 Cell source

The key element of 3D bioprinting is the bioink, which requires to meet some necessities, such as appropriate rheological, biocompatibility, and biological properties, structural cell growth support, and suitable mechanical properties.^{[180, 181]} The bioinks used in 3D bioprinting are categorized into two main groups: cell-scaffold-based approach and scaffold-free cell-based approach.^{[180, 182]} The cell-scaffold-based method uses a bioink that contains biodegradable biomaterial and living cells, and after the bioprinting process, biodegradation begins, and the cells grow to occupy the vacant space. In contrast, the scaffold-free cell-based technique bioprints the living cells directly, in a manner similar to normal embryonic growth. The source of cells used in bioinks also varies. The most commonly used cell sources for OOC applications include stem cells, primary cells, and immortal cells.

3.1 Heart and vessels

The engineering of cardiac tissues and organ models remains a great challenge, due to the special structure of the native myocardium, together with the need to integrate blood vessels, which adds to the complexity. Although in vivo models provide the appropriate environment in terms of physiology and biology, an in vitro surrogate is being introduced for research on tissue development and its functionalities through the advancement of 3D bioprinting technology, which is a reproducible and scalable fabrication methodology with precise 3D control compared with conventional tissue fabrication methods.^{[75, 213]} Moreover, scaffold-free 3D bioprinting can be integrated with conventional 3D tissue engineering methods to obtain a more realistic functional heart and advance science in the treatment of cardiovascular diseases.^{[213, 214]}

In this context, 3D bioprinting has been used to produce endothelialized myocardium.^[76] Using a multicomponent bioink and microfluidic technology, ECs were bioprinted directly into microfibrous hydrogel scaffolds. In combination with a specially designed microfluidic perfusion bioreactor, the resulting endothelialized myocardium-on-a-chip system was espoused to demonstrate the cardiovascular toxicity of pharmaceutical compounds. Such a strategy could be applied to human cardiomyocytes derived from induced pluripotent stem cells to construct endothelialized human myocardium. The final diameter of the resulting microfibers after extrusion was 150 μm. Moreover, human umbilical vein ECs (HUVECs), homogeneously distributed after bioprinting, gradually formed a confluent endothelial layer around the microfibers (in 14 days), which resembled the pattern of blood vessel walls. It was also reported that the density of the adherent cardiomyocytes, immediately after bioprinting, was independent of the aspect ratios of the unit grid of the scaffold. It was indicated that perfusion of the bioreactor at a rate of 50 μL/min reduced the number of dead bioprinted cardiac microtissues by approximately 400%. Doxorubicin, an anticancer drug, was used for drug screening assays, resulting in a 70.5 and 1.62% (near to 0 bpm) decrease in cardiomyocytes beating rate, 6 days after exposure to 10 and 100 μM of doxorubicin, respectively, whereas control endothelialized myocardial organoids maintained 88.3% beating rate, highlighting the efficacy of the proposed heart-on-chip platform for drug analysis.^[76]

To mimic blood flow and model the in vivo structure of the vessel, GelMA was used to bioprint a 3D vessel-on-a-chip platform with ECs and smooth muscle cells (SMCs) on a microfluidic chip with cell viability of ∼90% (94, 91, and 88% at 1, 4, and 7 days post-printing, respectively).^[215] The microfluidic chip was carved out of polymethylmethacrylate (PMMA) using a high-precision computer numerical control engraving machine. Compared with the conventional culture platforms, the EC-SMC coculture chip model resulted in a greater upregulation of alpha smooth muscle actin (αSMA) and SM22 protein expressions of the SMCs, and maintenance of the SMC contractile phenotype, mimicking the microenvironment of the natural vasculature under fluid flow conditions. This method enabled the establishment of an in vitro vascular model for physiologically relevant studies and the exploration of pathological processes in the vessel wall.^[215]

Thrombosis and its complications are one of the major causes of morbidity and mortality from cardiovascular disease, bringing about more deaths than trauma and cancer combined.^[216] To study thrombosis more effectively, an in vitro thrombosis-on-a-chip was fabricated using sacrificial 3D bioprinting technology to explore potential therapies and understand cellular interactions.^[216] The sacrificial layer was printed as a scaffold for UV curation of GelMA hydrogel to form hollow microchannels of the model, which were populated with HUVECs, into which human whole blood was infused immediately after thrombosis activation. Crosslinking of the hydrogel with longer UV exposure times resulted in matrice with higher moduli (0.8 kPa for 25 s, and 0.65 kPa for 10 s). However, increasing the exposure time to 20 and 25 s amplified the death rate of the embedded cells in the GelMA hydrogels and limited the spread of the encapsulated fibroblasts, whereas 10 and 15 s of UV exposure caused little damage to the cells (>80% cell viability). The platform was able to mimic thrombus within 10 min after application of 0.1 M of CaCl₂ in Dulbecco's phosphate-buffered saline. The thrombolysis test was performed by forming an artificial thrombosis on this setup to investigate the clot-dissolution ability of the tissue plasminogen activator (tPA), which showed the efficacy of tPA treatment within 2 h. The platform allowed the simulation of blood flow in the human leg (the common site of thrombosis formation) for burst pressure (∼0.16 kPa) with different flow rates (ranging from 0.6 to 3 mL/h) and velocities (ranging from 0.19 and 0.54 mm/s). Potentially, specific patient-derived cells can be used in the fabrication process to study fibrosis pathology and personalized medicine for vascular fibrotic diseases.^[216]

A multimaterial cardiac microphysiological device was developed on a chip for direct noninvasive electronic readout of contractile stresses to perform dose–response studies of drugs that influence contraction strength or beat rate.^[217] The wells were incubated with fibronectin (FN) solution in PBS for 1 h, following well seeding by either human-induced pluripotent stem cell-derived cardiomyocytes (hiPS-CMs) or primary neonatal rat ventricular myocytes. Following culture, the sarcomere packing density (a measure of periodic sarcomere organization) increased from 0.07 to 0.2, and the sarcomere orientational order parameter increased from 0.11 to 0.32, from day 2 to day 14, demonstrating that hiPS-CMs undergo myofibrillogenesis and sarcomerogenesis during culture. Besides, 4 weeks after culture, a significant increase in sarcomere length from 1.7 to 1.8 μm was observed (from day 14 to day 28), indicating a more mature tissue. This platform showed the potency of on-chip platforms for the engineering of laminar cardiac tissues with a range of ordered architectures.^[217]

3.2 Brain and BBB

Brain tumors exhibit a dynamic complexity consisting of different cell types.^[218] Glioma is the most aggressive brain tumor known to be responsible for approximately 48.3% of malignant brain tumors and other central nervous system (CNS) tumors^[219] and causes nearly 3% of cancer-related deaths annually.^{[220, 221]} Although mouse models are widely used for large-scale genomic analyzes and the study of biological mechanisms of tumorigenesis, genetic differences between humans and mice lead to an erroneous recapitulation of human pathophysiology. Despite advances in multimodal medical treatment, ∼70% of grade-II glioma tumors still progress to grades III and IV, which are lethal within 12–14 months.^[222] To promote patient survival, low-cost and accurate platforms are needed to critically analyze the molecular biology of glioma cells, as well as their interactions with the immune system and potential drug candidates.^[223] 3D bioprinting is one of the emerging methods that can be used to fabricate brain-on-chips,^[224] especially for glioma modeling,^[104] along with other applications such as developing patient-specific therapeutics^[225] and studying drug resistance of brain tumor cells,^[226] which is also important for developing methods to prevent tumors recurrence.^{[227, 228]} In addition, some specific methods such as lab-on-a-printer technology can also be used for 3D bioprinting of neural tissues.^[229]

Glioblastoma (GBM), a grade-IV astrocytoma, is a rapidly growing and aggressive brain tumor. Identification of patient-specific drug sensitivity and the development of more beneficial personalized cancer therapies can be achieved by ex vivo GBM-on-a-chip models.^{[230, 231]} To bioprint an ex vivo platform onto a glass substrate, silicon ink (to print the chamber wall) and GBM cells as well as vascular cells (HUVECs) were embedded into brain-decellularized ECM (BdECM) or collagen gel to determine patient-specific therapy resistances and to investigate drug combinations for a more effective tumor-killing performance of treatments (Figure 3A).^[232] This platform was equipped with an oxygen gradient-generating system to create an in vivo-like biomimetic oxygen gradient in the chip (Figure 3B). The printing process was performed using an in-house 3D printer (with a tapered nozzle) to print silicone ink where the custom bioink was pushed down through a flat needle at a speed of 500 nl/s. Human GBM cells were embedded in the BdECM and collagen gels and showed >90% cell viability in both, with a higher cell proliferation rate in BdECM after 10 days (Figure 3C). Moreover, higher expression levels of genes encoding proangiogenic factors (interleukin 8 (IL-8) and vascular endothelial growth factor A) and ECM-remodeling proteins (matrix metallopeptidase 9 (MMP9), MMP2, MMP1, FN, and protein tyrosine kinase 2) were detected for the BdECM gel compared with the collagen gel after 3 days. After 14 days, the cluster of differentiation 31-positive (CD31+) ECs produced more active tubule networks in the BdECM gel compared with than in the collagen gel, while the BdECM gel also showed superior capacity in terms of angiogenesis of HUVECs. Adjuvant concurrent chemoradiation therapy (CCRT) with temozolomide (TMZ), the most common therapy for GBM, was applied on the chips prepared from the patients’ GBM cells. According to the current clinical approaches, patients in group X had the highest survival rate, and those in group Z had the lowest survival rate. Following drug treatment, the chips made from group X's cells showed <40% cancer cell viability in response to drug treatment, whereas the group Z GBM-on-chip models showed >60% cancer cell viability, indicating the ability of this platform to accurately predict malignancy grade (Figure 3D).^[232]

A custom-made inkjet bioprinter was used to bioprint sodium alginate as a matrix for encapsulating cells (HepG2 and human glioma cell line (U251)) onto PDMS microchannels (prepared by soft lithography) for drug stimulation and diffusion experiments.^[130] It was highlighted that higher concentration and viscosity of the hydrogel resulted in a stronger hydrogel structure, while making the bioprinting process more difficult (droplets did not detach from the nozzle). Using larger nozzles facilitated the printing of viscous materials while sacrificing ultimate print resolution. In addition, applying a higher voltage solved the detachment issue by increasing the force generated by the piezoelectric material at the nozzle head, while compromising cell viability. Therefore, there was a trade-off between the viscosity of the bioprinting material, the applied voltage, and the nozzle size. In the study, a 0.5% concentration of alginate and a voltage of 40 V were chosen to produce the proposed chip using an inkjet bioprinter. Higher hydrophilicity of the glass substrate improved cell adhesion to the substrate and print resolution. By optimizing the hydrophilicity, a resolution of 400–1000 μm was achieved. To evaluate the drug modeling capability, the prodrug Tegafur was used as a model drug that can be metabolized by liver cells into the anticancer drug 5-fluorouracil, which acts on the U251 cells (glioma cells). Tegafur concentrations of 100 and 1000 μM were applied and successfully reduced the cancer cell viability of U251 cells to ∼80 and ∼50%, respectively. In addition, the OOC platform clarified that Tegafur inhibits cancer cell proliferation and is effective on U251 cells in the copresence of HepG2 cells, whereas it is not effective in the absence of HepG2.^[130]

The BBB—the barrier of ECs with high selectivity that shields the brain and CNS from solutes in circulating blood—is one of the challenges in delivering therapeutics to brain cancer cells, especially the GBM tumor foci,^{[233, 234]} and hence modeling the BBB will be useful for cancer studies. The TPP technology was used to bioprint a 1:1 scale microfluidic BBB chip composed of porous microcapillaries.^[235] Inspired by brain capillaries, the microtubes of the system were designed and bEnd.3 ECs from mouse brain and U87 GBM cells were used for seeding inside the chip. Numerical analyzes including intensity plots of fluid velocity and axial velocity profiles were used to investigate the proposed model in terms of the flow occurring, which demonstrated a uniform flow rate inside the channels similar to real physiological conditions. Confocal laser scanning microscopy 3D imaging of the model depicted that the bEnd.3 ECs were able to efficiently cover the tubular structures after 3 days of culturing. Furthermore, high-magnification scanning electron microscopy imaging of the setup provided qualitative evidence that the ECs almost completely covered the pores. By modifying parameters such as the diameters of the pores and microcapillaries, the pore density, and the length of the porous segment, the platform could be used as an in vitro model for a variety of studies such as drug screening.

Screening of drug permeability in the BBB was also investigated using a microfluidic BBB-on-a-chip fabricated with an SLA 3D printer.^[81] The fabricated pumpless chip consisted of three layers (lid, chamber, and perfusion), as shown in Figure 4A, and was able to mimic the in vivo properties of the BBB, making it a suitable platform for in vitro investigation of drug permeability. Brain microvascular ECs, derived from hiPSCs, were used for this study. These cells were cocultured for up to 10 days with rat primary astrocytes in a microfluidic system (Figure 4B), designed based on the residence time of blood in human brain tissues to ensure physiological transfer of nutrients and exogenous substances in a realistic manner around the BMECs without the need for pumps or tubing in the microfluidic system. The amounts of trans-endothelial electrical resistance (TEER) in the studied system were similar to the in vivo values. The TEER value was ∼3500 Ω · cm² and decreased to as low as ∼800 Ω · cm² after 24 h treatment with 0–10 μm doxorubicin (Figure 4C), demonstrating the achievement of significant barrier integrity in the proposed microfluidic system.

A 3D vascularized neural construct was developed for in vitro reconstitution of BBB function.^[236] 3D interconnected blood vessels were simulated by seeding ECs within the channels of the network, along with other types of cells, including neurons, astrocytes, and pericytes, in a collagen matrix, wrapping the vasculature network to derive a vascularized neural construct that recapitulates in vivo BBB function. A dopamine/collagen coating was used on the surface of the PCL/PLGA tube for the functionalization, increasing the surface hydrophilicity (water contact angle reduced from 104.7° to 33.8°) for improved cell adhesion and growth. Successful maturation of the culture was examined with the TEER test (raising from ∼40 to ∼120 Ω · cm² after 8 days), indicating a matured vascularization process for restricting compound transportation. The barrier capacity of the developed culture was tested by measuring the leakage of fluorescently labeled dextran of different molecular weights (10 kDa (2 nm), 70 kDa (11 nm), and 150 kDa (15 nm)). The cross-endothelial spreading in the peri-tubule region was 13% for 10 kDa (2 nm) dextran after 15 min, which was 40% for even larger 70 kDa (11 nm) dextran in the nonendothelialized case, showing the favorable impermeability of the developed model. Cell viability of ∼80% was recorded for the perfused model, compared with 50% in the nonperfused chip. Moreover, compared with nonvascularized tissue, neuronal growth in the vascularized tissue construct was longer with continuous neurite extension (perfused: 175 μm; nonperfused: 117 μm), underlining the vital supporting role of the vasculature system in the reconstruction of functional in vitro models.^[236]

In another study, a fluorescent molecule (fluorescein isothiocyanate (FITC)-dextran)-based test was used to investigate receptor-mediated transcytosis for screening purposes (receptor-mediated transcytosis is one of the major routes for drug delivery of large molecules into the brain).^[237] In this regard, immortalized human brain ECs, astrocytes, and pericytes were used to form a blood microvessel, which was grown adjacent to ECM gel, and a third channel to insert astrocytes and pericytes. The lumen of the endothelial vessel of the model was perfused with a controlled antibody or an antitransferrin receptor antibody in order to test antibody transcytosis. The barrier function/integrity was assessed using fluorescent barrier assay in which the leakage of a 0.1-mg/mL 20 kDa FITC-dextran dye (with a hydrodynamic radius of 3 nm) from the microvessel into the adjacent gel channel was measured by the acquisition of fluorescent images over time (using an ImageXpress XLS Micro HCI system (molecular devices)). Penetration of the antibody targeting the human transferrin receptor (MEM-189) was markedly higher (permeability of 2.9 × 10^–5 cm/min) than penetration of the control antibody (1.6 × 10^–5 cm/min), showing the potency of the developed OOC platform to retain all fluorescent dye within the vessel.^[237]

3.3 Lung and airway

According to the World Health Organization, chronic respiratory diseases (CRDs) (i.e., lung and airways diseases) are currently not completely curable. Nonetheless, drug treatments can control symptoms and relieve breathlessness by dilating the airways to improve patients’ quality of life.^[238] CRDs include asthma (235 million people suffer from it, mostly affecting children, 14% of children worldwide^[238]), lung cancer (the leading cause of cancer-related deaths (1.8 million deaths) with 2.2 million newly diagnosed cases in 2020^[239]), chronic obstructive pulmonary disease (causing 3.23 million deaths in 2019^[240]), pulmonary hypertension (occurs in ∼1% of the world's population), and occupational lung disease.^[241] Severe air pollution, smoking, occupational chemicals, and childhood lower respiratory tract infections are important risk factors contributing to CRDs. Lung-on-a-chip platforms allow correlation of various risk factors (e.g., genetics and chemicals) with the likelihood of developing CRDs, as well as real-time analysis of drug responses to patient-derived cells to achieve more effective personalized medicine and treatment. In recent decades, there have been attempts to fabricate airway-on-a-chip platforms.^{[242, 243]} For instance, a porous PDMS membrane was used to simulate the alveolar–capillary interface of a human lung on a chip to model the disease of pulmonary edema and to study organ-level responses to bacterial and inflammatory cytokines.^[243] However, this model faced the challenge of stably reproducing the dynamic structures of native 3D vascular networks over the long term. In addition, most conventional OOC platforms required manual fabrication steps that limited reproducibility.^[243]

An airway-on-a-chip, with an integrated natural vascular network, was 3D bioprinted in vitro using decellularized ECM bioink derived from porcine tracheal mucosa-derived dECM (tmdECM), as bioink, on a PCL frame to simulate respiratory disorders, such as allergen-induced asthma exacerbation and asthmatic airway inflammation (Figures 5A and 5B).^[242] The tmdECM hydrogel was prepared from the porcine trachea, where its rheological properties were studied before and after incubation and compared with collagen I (Col-1) and Matrigel using an advanced volumetric expansion system. The vascular platform, including PCL, EC bioink, and lung fibroblast bioink, was bioprinted using an in-house pneumatic extrusion-based 3D bioprinter. The fabricated platform with a mixture of tmdECM and Matrigel was able to form an extensive vascular network compared with Matrigel alone (Figure 5C). Self-assembly of ECs in tmdECM resulted in an interconnected vascular network after 7 days that was stable for up to 3 months under static culture conditions in vitro. The VP was also useful for monitoring tissue homeostasis, as its TEER (an indicator of cellular barrier integrity in the airway epithelium) was within the range of a healthy human tracheal epithelium. The vascularized airway-on-a-chip (VA-OC) was treated with IL-13, a cytokine that induces allergic asthma in human airway epithelium, to test the chip as an in vitro asthma model. Without affecting cell viability, treatment triggered the production of hRANTES (human regulated upon activation, normal T cell expressed and presumably secreted) and human tumor necrosis factor-α, the major inflammatory cytokines in asthma, at rates of 13.327 and 15.479 pg/mL, respectively, demonstrating the reproducibility of pathological interactions between the airway epithelium and vascular network using the VA-OC platform.^[242]

In order to recapitulate key features of the lower respiratory airways on a chip (specifically the blood vessel–interstitium fibroblast–epithelial microenvironment), a thin film (polyester track-etched (PETE) or vitrified collagen) membrane and a microvascular platform were integrated to grow epithelium at an air–liquid interface.^[244] The macrophysiologic device was comprised of a commercially available bottomless 96-well plate, PETE or vitrified collagen membranes, a 3D printed airway layer, and a photolithography-based vascular layer. Normal human lung fibroblasts (NHLF), in a fibrin gel, were seeded into the two outer channels, which provided the needed cytokine gradient to direct vasculogenesis of the HUVEC fibrin gel culture in the central channel. The PDMS layer was designed for epithelial cell growth and providing the open ports to seed and feed both culture layers. For NHLF-B devices, the αSMA median area expanded from 0.4% (for the untreated group) to 1.9 and 1.0% for pirfenidone and TGF-β1 treatment, respectively. In addition, the fluorescent area of the migrated neutrophils across the vasculature was higher in the PBS cystic fibrosis human bronchial/epithelial cells (PBS CF-HBE) devices compared with the normal human small airway epithelial cells, confirming that the device accurately modeled how CF epithelial cells increase neutrophil migration compared with normal epithelial cells. Despite the successful implementation of the model, it was suffering from a lack of expansion/contraction of the membrane to replicate the cyclic mechanical strain of breathing, and a lack of dynamic movement of media in the endothelial compartment.^[244]

3.4 Liver

The liver is the main organ for protein synthesis, bile acid production, biotransformation of drugs, detoxification, and filtration in the body.^[245] Since administered drugs enter the liver via the bloodstream, drug-induced liver injury is prevalent, highlighting the unmet need for the development of in vitro liver models for preclinical and clinical drug screening. Liver-on-a-chip methods are beneficial for understanding the liver function and disease impairment, studying the influences of dietary and cosmetics supplements on the human organ, screening foodborne pathogens/diseases, and monitoring drug toxicity/efficacy.^{[29, 246]} Despite the high proliferation and regeneration capacity in conventional 2D and 3D cultures, the functionality (e.g., enzymatic activities and albumin secretion) of in vitro hepatocytes from human liver biopsies decreases after being removed from the natural environment of the liver (within 1 week)^[247] because of the lack of an efficient perfusion system for waste product removal and nutrient delivery. On the other hand, liver cells can be cultured on a chip for a longer period of time in that liver-on-a-chip platforms provide continuous perfusion through microfluidic channels.^[77] Cell sources for liver-on-a-chip platforms include stem cell-derived hepatocytes (a consistent source of hepatocytes, but requiring specific induction factors and facing difficulties to manipulate), liver-derived cell lines (easy to manipulate and endowed with unlimited lifespan, but with rapid loss of expression of liver-specific transporters/enzymes and inaccurate drug response), and primary human hepatocytes (possess intrinsic properties of the liver, but are unsuitable for long-term culture, costly, and difficult to isolate).^{[246, 248]}

For analysis of human HepG2/C3A spheroids for drug toxicity over the long term (30 days), a continuously perfused, syringe pump-operated liver-on-a-chip setup was bioprinted (Figures 6A and 6B).^[77] PDMS was cast around a laser-cut PMMA mold to form a set of channels and three chambers. Unlike most PDMS-based microfluidic chips, where the different layers are permanently sealed by plasma, this setup was sealed with screws and nuts to prevent compound leakage, facilitating disassembly of the chip at any stage of the experiment for direct cell culture analysis. Fifteen seconds of UV light exposure (850 mW) was used to crosslink the bioink that were consisted of GelMA and HepG2/C3A spheroids (191 ± 10 μm). Drug simulation was performed with 15-mM acetaminophen (APAP) to induce a toxic response in the hepatic culture, giving results comparable to those obtained in animal experiments. The cell number was increased tenfold after 30 days of cell culture, which is a common cell density range for hydrogel encapsulation. By measuring the secretion of biomarkers (albumin, ceruloplasmin, alpha-1 antitrypsin (A1AT), and transferrin), the hepatic function of the proposed setup was examined.^[77] Similar setups were proposed, which were able to efficiently recapitulate key hepatic tissue cell types, such as hepatocytes, Kupffer, endothelial, and stellate cells, the distribution of various ECM-like biomaterials, and their ratios.^[249]

In another study, a gravity-induced (i.e., pumpless) liver-on-a-chip setup (0.03 × 0.1 × 1 mm³) was 3D bioprinted in a single step with perfusion capability (at a flow rate of 25 μL/min). The chip consisted of vascular/biliary channels for waste removal (lower biliary channel) and supplying nutrients (upper vascular channel), two reservoirs to ensure a continuous supply of compounds within the channels, and a 3D liver-decellularized ECM (dECM) environment for the cells (Figures 6C and 6D).^[245] Structural printing was performed under 660 kPa and 110°C conditions using poly(ethylene/vinyl acetate) and sterilized transparent PMMA as structural material and printing substrate, respectively. Cells were bioprinted using gelatin and liver dECM bioinks. The precision of cell bioprinting was evaluated by acquiring images of the cell layer, showing that HUVECs were accurately placed over the human hepatoma (HepaRG) cell-laden liver dECM bioink (Figure 6D). To examine the ability of this platform in emulating the basic function of liver cells, albumin/urea secretion in 2D culture was compared with that of the liver-on-a-chip platform. This demonstrated a higher level of secreted urea and albumin in the 3D bioprinted model, whereas the secretion level in the 2D model continuously decreased. Furthermore, the drug response of the model was analyzed by applying 5-mM APAP and measuring albumin secretion, reporting a higher drug sensitivity and reactivity in the 3D-bioprinted platform compared with the 2D model. The simulation of the liver-like microenvironment and the presence of multiple cell types and biliary channels were the main reasons for promoting the functionality of this setup.^[245]

Despite the considerable attention given to OOC platforms, this technology still suffers from protein absorption, challenges in forming different types of ECM environments for cell–ECM interactions, and poor selectivity of different cell types in the presence of spatial heterogeneity, mainly due to conventional fabrication methods (e.g., soft lithography, microcontact printing, and molding). To overcome these limitations, PCL may be a practical candidate. PCL is a nontoxic, biodegradable, and biocompatible polyester with a low melting point (∼60°C), resulting in high cell viability for 3D bioprinting. A single-step 3D bioprinting approach was introduced, without a secondary cell seeding process, to produce a perfused liver-on-a-chip platform with PCL-based 3D printed microfluidic channels (Figures 7A and 7B).^[8] PCL, as microfluidic chip material, and hydrogels containing encapsulated HUVEC and HepG2 cell lines were printed using a pneumatically activated nozzle 3D printer with a channel size of 1.5 mm × 1.5 mm × 15 mm and minimum line widths of 175 μm. The protein absorption rates of the PCL-based bioprinted 3D channels (protein absorption depth in the channel wall: 50 μm, absorbed dye in the channel: 3.4% of inlet concentration) were compared with those of soft lithographic PDMS-based channels (absorption depth in the channel wall: 400 μm; absorbed dye in channel: 10.5% of inlet concentration), confirming the better performance of PCL-based channels on heterotypic cell types and biomaterials for more accurate drug screening. However, PCL-based microfluidic channels exhibited lower optical transparency compared with PDMS channels (Figure 7C), which is one of the important features for cell proliferation/viability rate assessment and real-time analysis of cell interactions and drug responses.^[8]

3.5 Gut

Colorectal cancer, as the second leading cause of cancer-related death in the United States, was diagnosed in 147,950 individuals, in 2020, resulting in 53,200 deaths.^[250] To find more effective therapies, an OOC-like platform was implemented on a PDMS-based microfluidic chip (molds made of Pluronic F127) through a microextrusion 3D bioprinter to study colorectal carcinoma (Figure 8A).^[2] The structure consisted of channels (allowing physiological fluid flow onto the cell constructs) with width, length, and depth of 800 μm, 30 mm, and 300 μm, respectively, and concave wells (to hold the 3D-HCT116 cell constructs) with a diameter of 1.5 mm. Bioink, consisting of alginate and nanofibrillar cellulose, was mixed with HCT116 cells (in a 10:1 ratio) and bioprinted onto concave wells through a 410-μm nozzle at 4 kPa. The presence of bioink not only provided an analogous microenvironment to native ECM in the human body, but also protected the cells during bioprinting through the nozzle, with enhanced cell viability. The viability of the constructs was examined using cell nuclear staining and a fluorescence microscope. The cell viability of the 3D spherical cultures was measured to be 80.1, 67.8, and 64.7% at 1, 4, and 7 days after bioprinting, respectively (Figure 8B). Drug toxicity assays were performed with 7-ethyl-10-hydroxycamptothecin (SN-38), used in colon cancer, on the chip with three HTC116 construct arrays with separate microchannels. While the control line had a cell viability of 90%, application of 20 and 200-μM SN38 resulted in cell viability of 57 and 48%, respectively, 48 h after drug treatment (Figures 8C and 8D).^[2] Additionally, the potential impacts of bioprinting parameters were also scrutinized. Although smaller nozzle sizes resulted in better 3D bioprinting resolutions, reducing the nozzle dimensions from 410 to 200 μm increased the pressure required for bioprinting from 90 to 110 kPa, which augments the shear stress that cells are subjected to during the bioprinting process, which may have a negative effect on the cell viability rate. Besides, changing the nozzle shape from a conical to a needle-shaped nozzle increased the minimum bioprinting pressure by ∼120% for the same size. Moreover, the prevention of entrapped air bubbles is a challenging problem in 3D bioprinters, which can be addressed by optimizing the bioprinting characteristic, such as bioprinting pressure as well as speed, nozzle size, composition, and proportion of constituents in the bioink.^[2]

3.6 Renal system

The human kidney, filtering nearly 180 L of blood daily, is susceptible to blood-borne diseases and drug-induced injuries.^[251] Chronic kidney disease (CKD) refers to the malfunction of the kidney in filtering blood. As the ninth leading cause of death in the United States, the economic burden of kidney diseases was roughly $118.4 billion in the United States, in 2018, including diagnosis, drug treatment, kidney transplantation, and dialysis.^[252] Hypertension, diabetes, a family history of kidney failure, and heart disease are the main risk factors for developing kidney disease—three out of four cases of kidney failure are due to diabetes or hypertension.^[253] CKD can result in kidney failure, heart disease, anemia (i.e., low red blood cell count), infection, low calcium, high potassium, and high phosphorus levels in the blood. It is estimated that 5.24 million people will require dialysis by 2030, highlighting the urgent need to develop effective treatments for kidney disease.^[254] Early kidney models, cultured on hollow fibers or biomimetic basement membrane coatings, were able to maintain a differentiated state and self-organization of cells.^[255-260] Later, more complex 3D microenvironments were proposed, such as differentiated proximal tubule cells in thin gels^{[261, 262]} and induced pluripotent stem cell-derived kidney organoids with nephronal properties.^[263-266] Kidney-on-a-chip platforms mimic the structure of the human kidney and its functions, such as reabsorption, and thus presenting a promising tool for disease simulation, drug screening, and personalized therapies. Current kidney-on-a-chip platforms face the challenge of emulating complex 3D structures, drug responses, and tissue functions, such as reabsorption^[251]—the reentry of amino acids, potassium, glucose, and water filtered during glomerular filtration into the bloodstream via passive concentration gradient-based transfer in the proximal convoluted tubule of the nephron.^[267] Despite achieving complex 3D structures, the kidney organoids faced challenges in longevity, perfusion under physiological shear stress, perfusate collection, and analysis similar to 3D convoluted and open luminal architecture.^{[72, 268]} The 3D bioprinting technology can produce complex luminal tissue architectures, overcoming the current limitations of kidney models.

Human renal proximal tubules were printed in 3D on a chip with a perfusable open lumen architecture, confined by proximal tubule epithelial cells (PTECs), and analogous physiological shear stresses that can be maintained viable for 2 months.^[72] Both the silicon gasket and the fugitive ink (38 wt% Pluronic F127 and 100 U/mL of thrombin in deionized ultra-filtered water), for the tubular hollows, were printed onto a glass slide using a pneumatically activated nozzle extrusion 3D printer. The fugitive ink was then evacuated, leaving hollows inside the ECM for proximal tubules. PTECs, designed for proliferation, were then seeded to form the tubules and cultured using a continuous supply of cell media at shear stresses between 0.1 and 0.5 dynes/cm². Using transmission electron microscopy, it was observed that the growth height of the 3D perfused tissue cells was 14.1 μm, which was 100% more than the growth height of the nonperfused planner platforms, 40% more than that of the perfused 2D culture, and closer to healthy human proximal tubules (20.3 μm). Furthermore, the average microvilli length on 3D perfused tissue (1.24 μm) was also 200% longer than that of 2D nonperfused culture, ∼40% more than that of 2D perfused platforms, and was closer to in vivo values (2.89 μm), demonstrating the superiority of 3D-bioprinted platforms, over other conventional cultures, in simulating in vivo-like conditions. Albumin uptake, a crucial indicator of homeostasis, was observed to be higher in 3D tissue than in 2D. Moreover, the expression of megalin, one of the transporters for albumin, was the highest in 3D-printed culture among 2D cultures. Finally, for drug tests, the effects of cyclosporine A (CysA), a harmful nephrotoxin for proximal tubule cells, were monitored by perfusing it into the cell culture at varying concentrations. The epithelial barrier permeability was increased sixfold and fourfold by exposure to 100 and 500 μM of CysA, respectively.^[72]

In another study, using a modified ECM, an on-chip perfusable 3D human vascularized proximal tubules tissue was 3D-printed using a tubular–vascular exchange to simulate renal reabsorption (Figures 9A–9C).^[251] Although a minimum channel diameter of 20 μm was printable, seeding the high density of cells into channels smaller than 200 μm was challenging. It was reported that reducing the gelatin-to-fibrin ratio from 7.5 to 0.4, in the modified ECM, reduced the time required to obtain a confluent epithelium (fourfold reduction from ∼21 days to ∼4 days). Besides, a drug screening assay was performed on the 3D-printed kidney-on-chip using dapagliflozin, a glucose reabsorption inhibitor. Application of dapagliflozin resulted in ∼98% reduction in glucose reabsorption, highlighting the specificity and regulability of the developed on-chip model for longitudinal studies of drug responses (Figure 9C). Hyperglycemia, a sign of diabetes and a risk factor for vascular disease, was also modeled on this platform by circulating a perfusate with a fourfold higher glucose level and monitoring EC damage.^[251]

3.7 Breast

Cases of breast cancer, the most commonly diagnosed cancer in women, increased by 0.3% per year from 2012 to 2016.^[269] However, the mortality rate of breast cancer was reduced by 40% (averting 375,900 deaths) from 1989 to 2017, largely due to the development of effective therapies.^[269] Still, a controlled culture of breast cancer cells and administration of drug candidates to monitor the effects of treatments on the cancer cell and surrounding tissue may contribute to more effective therapies. Nonetheless, drug screening and recapitulation of cancer cell growth in an in vivo-like 3D structure are still a challenge in the development of chemotherapy, along with surgery and radiotherapy, as the most common cancer therapy.^[17] In this regard, a drug screening system, termed 3D tumor array chip (3D-TAC), was bioprinted with GelMA hydrogel droplets (∼0.1 μL) containing MDA-MB-231 breast tumor cells and evaluated with epirubicin as well as paclitaxel, antitumor drugs, to demonstrate the compatibility of this platform with traditional screening approaches.^[17] Although the chip basement was a transparent conductive membrane, the culture chambers consisted of a silicon interlayer and stainless steel. MDA-MB-231 cells, encapsulated in GelMA, were bioprinted onto the conductive membrane using a high-voltage electrohydrodynamic 3D bioprinter through stainless steel nozzles at 2.7 kV and a flow rate of 10 μL/min, and crosslinked with 405-nm light. Despite the high electric field force during the process, the cell viability was above 90%, indicating the high performance of the electrohydrodynamic bioprinting technique. To investigate the proliferation behavior of MDA-MB-231 cells, the nicotinamide adenine dinucleotide (NAD+) examination was tested by the enzyme-linked immunosorbent assay (ELISA) method. The half-maximal inhibitory concentration (IC₅₀) values of the 3D models were 47.63 and 28.83 μM for paclitaxel and epirubicin, respectively, higher than those of the 2D at 46.09 and 27.77 μM, confirming the inhibition of tumor cell proliferation by the applied drugs.^[17]

Although the integration of blood vessels into OOC platforms for perfusion is widespread, there are a few reports of the simultaneous implementation of blood and lymphatic vessels in in vitro tumor models. The lymphatic system drains fluid (known as lymph) that has seeped into tissues from blood vessels and returns it to the bloodstream—it plays a key role in the immune system by protecting the body from disease-causing invaders, removing cellular waste, absorbing fats from the digestive tract, and maintaining the body's fluid balance.^[270-272] In cancer therapies, the lymphatic drainage system provides a preferential recycling pathway for applied antitumor drugs in vivo and plays a critical role in metastasis, highlighting the importance of developing perfusable OOC platforms with blood and lymphatic vessels.^{[18, 273-277]} A tumor-on-a-chip platform was bioprinted on a microfluidic bioreactor with a dynamic microenvironment consisting of blood and lymphatic vessel pair and MCF-7 breast cancer cells in a 3D hydrogel matrix based on GelMA on a PDMS–PMMA frame (Figures 10A–10F).^[18] The vessels were bioprinted using a custom-made bioprinter with a trilayered coaxial extrusion channel featuring different needle sizes. The permeability values of the vessels were measured by observation of FITC–bovine serum albumin and determined to be ∼5.07 × 10^–7 cm/s for the blood vessel and ∼1.76 × 10^–6 cm/s for the lymphatic vessel, which was comparable to the native in vivo values of vessels (1–8 × 10^–7 cm/s for blood vessels, 0.1–5 × 10^–6 cm/s for lymphatic vessels^[18]). It was also reported that the presence of a second channel (lymphatic vessel) not only caused higher cell viability (Figure 10G) by avoiding drug accumulation in the cell culture, but also accelerated the diffusion rate of FITC compared with the simple perfusion case (blood vessel only) (Figures 10H and 10I). Furthermore, embedding cancer cells in the GelMA slowed the rate of drug or biomolecule (FITC) transfer.^[18]

3.8 Bone and cartilage

The study of bone and cartilage models not only helps to understand common bone defects and fractures, but also paves the way for a better understanding of various tumors such as lung, breast, prostate, and melanoma carcinomas, for which the skeletal system of the human body is a chronic site for metastasis of these cancer cells (for instance, in metastatic breast cancer, bone metastasis occurs with a frequency of ∼70%^[278]).^[279] Although the 5-year survival rate of breast cancer patients, without bone metastasis, can be as high as 98% with early diagnosis and therapy, the occurrence of bone metastasis decreases this rate to 26%, highlighting the importance of understanding the mechanism of bone metastasis and the means of prevention.^[278] In addition to cancer, mild bone diseases (e.g., osteoporosis and low bone mass) and subsequent complications affect the quality of life and cause socioeconomic impact. Approximately 43.6 million people over the age of 50 in the United States suffer from bone diseases, resulting in hospitalizations for hip, spine, or wrist fractures with an estimated health care burden of $25 billion in 2025.^{[280, 281]} Common challenges in bone analysis with current models include: the complexity of obtaining antemortem samples of bone metastases, a limited number of ex vivo and in vivo models that effectively emulate bone structure, the physical difficulty of manipulating bone as tissue, and the limitation imposed by examination times, making the real-time monitoring of bone diseases and bone metastases difficult.^[278] Recent advances in 3D bioprinting, microfluidics, and OOC science have led to a narrowing of the gap between laboratory constructs and physiological tissues, resulting in the introduction of personalized therapies,^[282] a better understanding of the underlying causes of bone diseases, and more effective reconstitution of the cancer microenvironment.^[283-287]

Using a desktop 3D printer to create plastic molds followed by replica molding with PDMS, a spontaneous 3D bone-on-a-chip was fabricated to study breast tumor metastasis.^[278] As illustrated in Figures 11A–11C, the bone-on-a-chip setup was fabricated based on simultaneous growth dialysis and contained a solution reservoir in the upper part and a cell growth chamber in the lower part. The design of the chip was able to allow long-term growth for up to 30 days, resulting in the formation of a thick mineralized osteoblastic tissue in vitro. To characterize the maturation of osteoblastic tissue on the chip, normalized alkaline phosphatase activity, over 720 h of culture, showed that nitrocellulose-bone-on-a-chip (NC-BC) had the highest activity (Figure 11D). In addition, normalized measurement of insoluble ECM demonstrated that NC-BC collected the highest amount of insoluble ECM (Figure 11E). Etching the glass coverslip with a base solution for a longer period of time (72 h instead of 24 h) increased surface nanoroughness and surface hydrophilicity, which ultimately improved osteoblastic tissue attachment for long-term cultures. In addition, reducing the height of the culture chamber from 5 to 2 mm promoted both mineral and collagen deposition. It was shown that several important features of breast cancer colonization in bone were observed when coculturing metastatic human breast cancer cells with the osteoblastic tissue using the fabricated bone-on-a-chip. For instance, the colony-forming potentials of MDA-MB-231GFP and MDA-MB-231-BRMS1GFP cells were compared, and the former was found to be more likely to form micrometastases (88.5%) than the latter (61.4%).^[278]

3.9 Skin

In 2018, nearly 1.3 million new cases of skin cancer were diagnosed, with 64,000 reported deaths from nonmelanoma skin cancer. Encouragingly, the mortality rate from melanoma skin cancer has steadily declined since 2013, decreasing by 6.4% per year, due to the introduction of new treatment strategies, such as targeted therapies for metastatic melanoma and immune checkpoint inhibitors, indicating the importance of developing new diagnostic and treatment modalities.^[239] In addition to cancer, skin lesions also pose a significant burden on the healthcare system. The global wound care market is expected to grow from $19.3 billion in 2021 to over $27.8 billion by 2026.^{[288, 289]} Moreover, the global tissue-engineered skin substitutes market is expected to reach $3.9 billion in 2023, with a CAGR of 17.2%.^[288] In recent decades, remarkable progress has been made in modeling and developing in vitro-engineered substitutes that mimic human skin^[290] for various purposes such as developing in vitro skin models,^[291] studying toxicology tests,^[292] and creating skin-on-a-chip models for drug testing.^[293] 3D bioprinting has further accelerated the relevant science in this field by enabling standardized reproducible fabrication methods using various bioinks^{[288, 292, 294]} or direct cell bioprinting.^[133] Although skin-on-a-chip devices^{[291, 293, 295-297]} and 3D-bioprinted skin platforms^[298-300] are available in the literature, to our knowledge there is no report on “3D-bioprinted skin-on-a-chip”. Future studies could focus on the development of 3D-bioprinted skin-on-chip to take advantage of both OOC platforms and 3D bioprinting, such as continuous perfusion, precise control of dimensions, and multicell cultures. The fabrication of a bioprinted skin begins with the collection and culturing of cells.^[290] Various cells such as keratinocytes, fibroblasts, and melanocytes are cultured and grown after being collected from the human body. The biopolymer matrix is then used to produce the bioink to be used in the bioprinter. Then, the CAD model is prepared depending on the final application. After bioprinting, the fabricated system must undergo different steps in vitro to be ready for further applications such as implantation.

3.10 Multiorgan platforms

The ultimate goal of OOC platforms is to produce multi-OOCs to investigate interactions of different organs more realistically while monitoring possible responses and negative side-effects of administered medicine. However, most of the current OOC platforms are optimized for a specific organ (e.g., particular perfusion rate and circulating medium), complicating the culture of different tissues on a single chip. In this regard, as a proof-of-concept design, a multi-OOC platform was developed using the DLP technology with separate compartments, which enabled the culture of three different types of tissues on a single chip with tissue-specific perfusion rates to study tissue–tissue interactions through interconnective microfluidic channels (1-mm channel radius).^[301] To validate the ability of the developed chip to produce different perfusion rates in each compartment independently, 10 and 1 mm³/min flow rates were successfully applied to the first and second wells, respectively, with a laminar flow profile at 0.72 mm/s flow velocity. In addition, the chip was able to generate different shear stresses in each well, recording a shear stress of 0.001 and 0.02 dynes/cm² on the surfaces of the membrane in chamber one and chamber two, respectively.^[301]