**3. Organ‐on‐chip devices: concept to application**

In this section, various state‐of‐art existing OOC platforms and their structural features, working principles, potential and feasibility for biomedical application are discussed. OOC devices can be defined as microfluidics systems for living cells culturing in continuously perfused, micrometersized chambers in order to model physiological functions of tissues and organs [66]. The prime objective of this emerging technique is to fabricate minimal functional units of an organ that recapitulate tissue‐ and organ‐level interactions. These devices have great potential for investigating basic mechanisms of organ physiology and are well suited for the study of biological phenomena that depends on tissue microarchitecture and perfusion and last for relatively short span (< I month). These chips often consist of featuring multiple, controllable parallel channels, splitting and merging channels, various pumps, valves and integrated electrical and biochemical sensors. Some kind of microenvironment stimuli derived from organ‐level functions can be applied to cells from certain organ.

#### **3.1. Basic working mechanism of OOC devices**

OOC systems are basically elaborated microengineered physiological systems that reconstitute the key features of specific human tissues and organs and their interactions as depicted in **Figure 10** [82, 83].

Key factors in OOC designing include the following:


Earlier, with 2D and 3D cell cultures, efforts were taken to control and regulate the cell growth, shape and other cellular events but due to lack of precise 3D environment, these models suffered with inaccuracy and reliability in recapitulating the issue‐ and organ‐specific systems [83]. But with the state‐of‐art OOC technology, new possibilities to create efficient *in vitro* models with organ‐specific microenvironments, tissue microarchitecture reconstruction, spatio‐temporalchemical gradients, tissue‐specific interfaces, crucial dynamic mechanical cues and biochemical signals [54, 84]. In this section, we describe recent progress in this field and currently reported OOC devices such as liver, kidney, intestine, kidney, heart, skin and blood vessels.

**Figure 10.** Representation of organ‐on‐chip device and concept of modeling, a complex microenvironment and their existing simulation of functional units [82].

OOC devices can be classified into three broad segments based on the working mechanisms: [82]


#### **3.2. Membrane‐based organ on‐chip devices**

Excellent optical transparency is prime advantage of PDMS that enabled real‐time monitoring of nitric oxide production and variation in pulmonary vascular resistance in a microfluidic

Moreover, control of cellular parameters is another important phenomenon in designing OOC devices and recent advances in microfabrication techniques have significant contribution toward efficient monitoring and control of cellular responses and study of broad array of physiological factors that wasn't possible with 3D static cultures. Electrical, chemical, me‐ chanical and optical probes for direct visualization and quantitative analysis of cellular biochemistry, gene expression, structure and mechanical responses also can be integrated into virtually any microfabricated cell culture devices and more relevant data can be obtained with

In this section, various state‐of‐art existing OOC platforms and their structural features, working principles, potential and feasibility for biomedical application are discussed. OOC devices can be defined as microfluidics systems for living cells culturing in continuously perfused, micrometersized chambers in order to model physiological functions of tissues and organs [66]. The prime objective of this emerging technique is to fabricate minimal functional units of an organ that recapitulate tissue‐ and organ‐level interactions. These devices have great potential for investigating basic mechanisms of organ physiology and are well suited for the study of biological phenomena that depends on tissue microarchitecture and perfusion and last for relatively short span (< I month). These chips often consist of featuring multiple, controllable parallel channels, splitting and merging channels, various pumps, valves and integrated electrical and biochemical sensors. Some kind of microenvironment stimuli derived

OOC systems are basically elaborated microengineered physiological systems that reconstitute the key features of specific human tissues and organs and their interactions as depicted in

**•** Fabrication of OOC devices start with identifying the key aspects of biochemical, mechanical environment of specific organ, including local factors from neighboring cells or tissues and

Earlier, with 2D and 3D cell cultures, efforts were taken to control and regulate the cell growth, shape and other cellular events but due to lack of precise 3D environment, these models suffered with inaccuracy and reliability in recapitulating the issue‐ and organ‐specific systems

**•** The final step is to measure the functional output parameters of the cultured cells.

model and cell morphology, tissue repair and reorganization. [79–81]

**3. Organ‐on‐chip devices: concept to application**

from organ‐level functions can be applied to cells from certain organ.

**3.1. Basic working mechanism of OOC devices**

Key factors in OOC designing include the following:

**Figure 10** [82, 83].

stretch of organ. [82].

these advanced OOC devices. [54, 66]

92 Lab-on-a-Chip Fabrication and Application

To study the drug response with respect to human biological barriers is a crucial step in drug discovery. Researchers developed 3D compartmentalization with membrane‐based multilayer compartments for mimicking biological barriers such as the blood‐brain barrier [85, 86, 99], the kidney transport barrier [87, 71], and the lung's alveolar‐capillary interface [88, 89] that can be considered a major breakthrough for biomedicine. In this segment, recent discoveries in membrane and muscular thin films to recapitulate the physiochemical interface and mechan‐ ical cues are described.

#### *3.2.1. Lung on a chip*

Lung is an important organ of respiratory system for the exchange of oxygen and carbon dioxide in blood stream. The elementary tissue unit of the lung is the layer of epithelial and endothelial cells over which the exchange of gases between air and blood takes place. The geometry of the lung tissue contains the epithelial‐endothelial interface, epithelium‐air interface, endothelium‐blood interface and periodic mechanical force with each respiratory cycle. Understanding of cell‐cell interactions, cell‐blood and cell‐gas flow is utmost necessary for drug discoveries and physiochemical research. Complex geometric and compositional structure of lung is the great barrier to enable straightforward manipulation and observation of cells.

Lung‐on‐chip is the microreplica of the lung on a microchip. This is used for nanotoxicology studies of various nanoparticles that are introduced into the air channels and to understand the pulmonary diseases where due to the formation of liquid plug that blocks small airways and obstruct gas flow in alveoli [89]. To understand the mechanism of liquid plug propagation and rupture, Huh et al. designed a microengineered system that consists of two PDMS chambers separated by thin polyester membrane with 400‐nm pores. This system mimicked an *in vivo* basement membrane for small airway epithelial cells (SAECs) attachment and growth.

**Figure 11.** Schematic of lung on‐chip system. (a) PDMS‐based membrane to mimic alveolar capillary barrier and a vac‐ uum based deformation controller. (b) Size variation of lung during inhalation. (c) Bonding and alignment of three lay‐ er PDMS devices [70].

Using this system, injurious response of SAECs to propagation and rupture of finite liquid plugs at an air‐liquid interface afflicted with surfactant deficiency was demonstrated [88]. Another report by Huh et al. designed an alveoli‐on‐chip having alveolar and the capillary interface. To mimic the breathing pattern, two chambers were constructed at the side through which air is pumped in at certain required pressure, continuous increase and decrease of the flow is done in order to accomplish the inhalation and exhalation pattern. A thin flexible layer of PDMS was used in the central chamber where coculturing of human alveolar epithelial cells and blood vessel wall cells on the opposite sides is done. The membrane stretches and relaxes according to the flow of air. The culture medium is pumped through the lower microchannel to mimic the blood flow and the sample is injected on the top layer that interacts with the alveolar epithelial cells as shown in **Figure 11** [70]. In another model to study alveolar cell complexities, Douville et al. put forth their system consisting of two compartments—alveolar chamber and actunation channel. These chambers were separated by a PDMS thin membrane to create both cyclic stretch and fluid mechanical stresses. This *in vitro* model successfully demonstrated the difference in morphological changes cells undergo when exposed to combine stresses as compared to cells exposed solely to cyclic stretch [90].

These inventions reconstituted the critical lung functions and can be applied for *in vivo* models in environmental toxins, absorption of aerosolized therapeutics and the safety and efficacy of new drugs. Such a tool may help accelerate pharmaceutical development by reducing the reliance on current models, in which testing a single substance can cost more than \$2 million [54, 66].

#### *3.2.2. Kidney on chip*

*3.2.1. Lung on a chip*

94 Lab-on-a-Chip Fabrication and Application

of cells.

growth.

er PDMS devices [70].

Lung is an important organ of respiratory system for the exchange of oxygen and carbon dioxide in blood stream. The elementary tissue unit of the lung is the layer of epithelial and endothelial cells over which the exchange of gases between air and blood takes place. The geometry of the lung tissue contains the epithelial‐endothelial interface, epithelium‐air interface, endothelium‐blood interface and periodic mechanical force with each respiratory cycle. Understanding of cell‐cell interactions, cell‐blood and cell‐gas flow is utmost necessary for drug discoveries and physiochemical research. Complex geometric and compositional structure of lung is the great barrier to enable straightforward manipulation and observation

Lung‐on‐chip is the microreplica of the lung on a microchip. This is used for nanotoxicology studies of various nanoparticles that are introduced into the air channels and to understand the pulmonary diseases where due to the formation of liquid plug that blocks small airways and obstruct gas flow in alveoli [89]. To understand the mechanism of liquid plug propagation and rupture, Huh et al. designed a microengineered system that consists of two PDMS chambers separated by thin polyester membrane with 400‐nm pores. This system mimicked an *in vivo* basement membrane for small airway epithelial cells (SAECs) attachment and

**Figure 11.** Schematic of lung on‐chip system. (a) PDMS‐based membrane to mimic alveolar capillary barrier and a vac‐ uum based deformation controller. (b) Size variation of lung during inhalation. (c) Bonding and alignment of three lay‐ The word kidney‐on‐chip suggests that the kidney is mimicked on a chip. Here, the renal cells or the nephrons are mimicked on the chip and this is used for checking the toxicity of drug and its screening. This model helped to know more about the filtration, reabsorption of the necessary molecules from the drug as kidney toxicity is a cause of concern during drug development [91]. Nephron is the basic unit of kidney and mainly consists of glomerulus, which acts as a filtering unit that helps in filtering unwanted toxic particle from the required molecules and helps in throwing out these unwanted molecules. Nephron's glomerulus, proximal convoluted tubule and loop of Henle are mimicked on the chip. As reported by Weinberg et al., an artificial nephron function with three components on a single chip was designed [92]. Jang at al developed an on‐chip kidney to reproduce cisplatin nephrotoxicity. Their device contained two compartments, where top channel mimicked urinary lumen and has fluid flow, whereas the bottom chamber imitate interstitial space filled with media. Kidney cells have less shear stress than endothelial or lung cells. This device was operated with 1 dyn/ cm2 of sheer stress [93]. A modified version of same device using human proximal tubular cells was also developed by the same group. The advantage of using proximal cells was there less sheer stress ∼0.2 dyn/cm2 that is similar to that of the living kidney tubules surrounding as shown in **Figure 12** [94]. Better understating of filtration pattern and absorption behavior that leads to toxicity was the prime aspect of this discovery.

**Figure 12.** (a) Nephron on a chip: Schematic of the chip with cross sections of three functional units named glomerulus, proximal convoluted tubule and loop of Henle, which are response for filtration, reabsorption and urea concentration, respectively. (b) Kidney reabsorption functions using a microfluidic chip comprising of an apical channel separated from a bottom channel by proximal tubular epithelial cells cultured ECM‐coated porous membrane [94].

#### *3.2.3. Blood‐brain barrier on chip*

To understand and treat neurological diseases, proper understanding of blood‐brain barrier (BBB) is utmost important. By definition, BBB is a unique selective barrier membrane that obstructs the passage of most exogenous compounds in blood to the central nervous system (CNS) while permeable for essential amino acids and nutrients. It is made primarily of three different cells: endothelial, pericytes and astrocytes, and the membrane is formed by firm junctions between endothelial cells that control compound permeability with high values of transendothelial electrical resistance (TEER) [82, 95]. Hatherell et al. designed a membrane‐ based system to replicate BBB by cultivating endothelial cells on the top side of a transwell membrane while cultivating astrocytes with or without pericytes on the opposite side [88]. However, due to low porosity and uneven pore distribution, this artificial membrane failed to recreate the close proximity to cell interaction. To address this issue, silicon nitride membrane was developed by Ma et al. to increase the direct contact between the cells and astrocytes [96]. Another report by Shayan et al. also demonstrated a considerable reduction of the flow resistance across a nanofabricated membrane with controlled pore size and low thickness (3 μm) and maintenance of metabolic activity and viability for at least 3 days [97]. A novel BBB *in vitro* model was developed by Brown et al. for efficient cell‐to‐cell communication between endothelial cells, pericytes, and astrocytes and independent perfusion with vascular chamber and brain chamber separated by a porous membrane (**Figure 13a**) [98]. Booth and Kim also developed a BBB that impersonated the dynamic cerebrovascular environment having fluid shear stress and a comparatively thin culture membrane of 10 μm (**Figure 13b**). This system has two components called lumenal and ablumenal on which endothelial and astrocytes were cultured to form the neurovascular unit [99].

**Figure 13.** Schematic view of the neurovascular unit (NVU) indicating major components, cell types and their spatial arrangement. (a) Illustration of key properties should be included in an effective *in vitro* microfluidic blood‐brain barri‐ er (μBBB) models (left). (b) Structure of microdevice consisting of two channels for astrocytes and endothelial cells cul‐ ture with electrodes for transendothelial electrical resistance (TEER) measurement [98, 99].

These novel systems are the promising tools of future due to their unique characteristics of feasible real time, TEER and selective permeability to study barrier function and delivery of drugs to CNS.

#### *3.2.4. Heart on chip*

**Figure 12.** (a) Nephron on a chip: Schematic of the chip with cross sections of three functional units named glomerulus, proximal convoluted tubule and loop of Henle, which are response for filtration, reabsorption and urea concentration, respectively. (b) Kidney reabsorption functions using a microfluidic chip comprising of an apical channel separated

To understand and treat neurological diseases, proper understanding of blood‐brain barrier (BBB) is utmost important. By definition, BBB is a unique selective barrier membrane that obstructs the passage of most exogenous compounds in blood to the central nervous system (CNS) while permeable for essential amino acids and nutrients. It is made primarily of three different cells: endothelial, pericytes and astrocytes, and the membrane is formed by firm junctions between endothelial cells that control compound permeability with high values of transendothelial electrical resistance (TEER) [82, 95]. Hatherell et al. designed a membrane‐ based system to replicate BBB by cultivating endothelial cells on the top side of a transwell membrane while cultivating astrocytes with or without pericytes on the opposite side [88]. However, due to low porosity and uneven pore distribution, this artificial membrane failed to recreate the close proximity to cell interaction. To address this issue, silicon nitride membrane was developed by Ma et al. to increase the direct contact between the cells and astrocytes [96]. Another report by Shayan et al. also demonstrated a considerable reduction of the flow resistance across a nanofabricated membrane with controlled pore size and low thickness (3 μm) and maintenance of metabolic activity and viability for at least 3 days [97]. A novel BBB *in vitro* model was developed by Brown et al. for efficient cell‐to‐cell communication between endothelial cells, pericytes, and astrocytes and independent perfusion with vascular chamber and brain chamber separated by a porous membrane (**Figure 13a**) [98]. Booth and Kim also developed a BBB that impersonated the dynamic cerebrovascular environment having fluid shear stress and a comparatively thin culture membrane of 10 μm (**Figure 13b**). This system has two components called lumenal and ablumenal on which endothelial and astrocytes were

from a bottom channel by proximal tubular epithelial cells cultured ECM‐coated porous membrane [94].

*3.2.3. Blood‐brain barrier on chip*

96 Lab-on-a-Chip Fabrication and Application

cultured to form the neurovascular unit [99].

Heart on chip was developed to imitate the contractility and electrophysiological response of heart in *in vitro* condition. Microfluidics has previous applications *in vitro* on cardiomyocytes, which generates the electrical impulse that controls the heart rate. However, these previous experiments could not fully reconstruct the tissue microenvironments, such as the propagation of an action potential (AP) or generation of contractions. To fulfill specific needs of heart‐on‐ chip studies, a biohybrid construct was designed based on muscular thin films (MTFs); a tissue‐ engineered myocardium consist of anisotropic cardiomyocytes cultured on a deformable elastic thin film with various geometries [100, 101].

Grosberg et al. was pioneer in developing MTF‐based "heart‐on‐a‐chip" system that success‐ fully measured the contractility of neonatal rat ventricular cardiomyocytes exposed to various doses of epinephrine [100]. Eight separate MTFs were framed the skeleton of their system and was fabricated in batches enabling them to collect data from multiple tissues simultaneously in the same experiment. This heart‐on‐chip system mimicked the hierarchical tissue architec‐ ture of laminar cardiac muscle, and measurements of structure‐function relationships, including contractility, AP propagation and cytoskeletal architecture. In another approach, Agarwal et al. explored an optimized semiautomated microdevice to test the positive inotropic effect of different dosages of isoproterenol on cardiac muscle contractility. They achieved an increased drug‐screening throughput with their device having 35 separate thin films (**Figure 14a**) [102]. Basic components of this device includes a semiautomatic microdevice integrated an MTF chip, an electrode for electric field simulation, a metallic base on a heating element as temperature control unit and a transparent window for cantilever deformation monitoring. As these models were based on animal tissues and cannot recapitulate human system with precision. To overcome this limitation, Mathur et al. designed cardiac microphy‐ siological system (MPS) that could imitate the human myocardium and envisage the cardio‐ toxicity of drugs accurately, by merging hiPSC‐derived (human‐induced pluripotent stem cells (hiPSC)) cardiomyocytes with an appropriate microarchitecture and "tissue‐like" drug gradients (**Figure 14b**) [103]. These hiPSC‐derived cardiac MPS predicted drug response and toxicity *in vitro* and showed a wide applicability for disease modeling and drug screening [82, 103]. Few reports are also available to tackle this complex yet vital organ of our system [104].

**Figure 14.** (a) Graphical illustration of the fabrication process flow for muscular thin film (MTF) and the semiautomatic microdevice integrated a MTF chip [102]. (b) Schematic of the microphysiological system (MPS) with nutrient channels (red), cell‐loading channel (green) and 2 μm endothelial‐like barriers. Optical and confocal fluorescence imaging of 3D cardiac tissue aligned with multiple hiPSC cardiac cells layer [103].

#### *3.2.5. Stem cells on chip*

Human stem cells are a critical component for OOC devices. Few reports are available where stems cells were grown in scaffolds and microarrays. These controlled conditions make it possible to mimic the complex structures and cellular interactions within and between different cell types and organs *in vivo* and keep the culture viable over long periods of time. It was reported that neurogenesis of human mesenchymal stem cells can occur in the absence of chemical stimuli, simply through the substrate stiffness [105]. **Figure 15a** is illustrating a PDMS membrane‐based platform for stem cells growth.

In principle, all cell sources, whether primary cells (directly taken from an organ or tissue, e.g., by means of a biopsy needle), or cells or in the form of cell lines, from animal or human origin, can be useful for the OOC approach. The basic criterion for selecting the stem cells for OOC is target disease. For the diseases with well‐known gene mutation, the DNA, specific disease‐ causing DNA mutation can be introduced into a stem cell line by the technique of homologous recombination, resulting in two human cell lines with one having disease‐causing DNA mutation in one of them. For this purpose, both hES (human stem cells) and iPSC (induced pluripotent stem cells) sources can in principle be used. On the other hand, for diseases caused by a whole spectrum of mutations in any part of the disease‐causing gene, or diseases associated with a more complex genetic background, iPSC cell line or adult stem cells derived from a patient with the disease need to be used to recreate "the patient"—on a chip. iPSC cells are the first choice in contrast to adult stem cells, due to ease of regeneration [106].

**Figure 15.** (a) PDMS based on‐chip platform for stem cells. (b) Crypt‐villus structures grown from single LGR5 positive adult stem cells from the intestinal crypt [106].

Three‐dimensional "organoid" stem‐cell culture technology was developed in the laboratory of Hans Clevers at the Hubrecht Institute. In this approach, intestinal stem cells were isolated from the intestinal epithelial tissue by separating tissue cells from each other. Subsequently, few stem cells within the cell mixture were identified by coupling them to a specific fluorescent antibody, followed by isolation with a fluorescence‐activated cell sorter. 3D environment was created by the gel surrounding the cells to make them feel comfortable in their new "niche." In this process of cell growth, the stem cells were bound to their "mate," which is necessary to provide the essential cell‐cell contact to start the self‐renewal process. Once in the dish, each cell combination starts to self‐assemble, a new crypt‐villus structure in three dimensions forms called organoids (**Figure 15b**).

#### **3.3. Anatomy‐based organ function mimicking**

increased drug‐screening throughput with their device having 35 separate thin films (**Figure 14a**) [102]. Basic components of this device includes a semiautomatic microdevice integrated an MTF chip, an electrode for electric field simulation, a metallic base on a heating element as temperature control unit and a transparent window for cantilever deformation monitoring. As these models were based on animal tissues and cannot recapitulate human system with precision. To overcome this limitation, Mathur et al. designed cardiac microphy‐ siological system (MPS) that could imitate the human myocardium and envisage the cardio‐ toxicity of drugs accurately, by merging hiPSC‐derived (human‐induced pluripotent stem cells (hiPSC)) cardiomyocytes with an appropriate microarchitecture and "tissue‐like" drug gradients (**Figure 14b**) [103]. These hiPSC‐derived cardiac MPS predicted drug response and toxicity *in vitro* and showed a wide applicability for disease modeling and drug screening [82, 103]. Few reports are also available to tackle this complex yet vital organ of our system [104].

**Figure 14.** (a) Graphical illustration of the fabrication process flow for muscular thin film (MTF) and the semiautomatic microdevice integrated a MTF chip [102]. (b) Schematic of the microphysiological system (MPS) with nutrient channels (red), cell‐loading channel (green) and 2 μm endothelial‐like barriers. Optical and confocal fluorescence imaging of 3D

Human stem cells are a critical component for OOC devices. Few reports are available where stems cells were grown in scaffolds and microarrays. These controlled conditions make it possible to mimic the complex structures and cellular interactions within and between different cell types and organs *in vivo* and keep the culture viable over long periods of time. It was reported that neurogenesis of human mesenchymal stem cells can occur in the absence of chemical stimuli, simply through the substrate stiffness [105]. **Figure 15a** is illustrating a PDMS

In principle, all cell sources, whether primary cells (directly taken from an organ or tissue, e.g., by means of a biopsy needle), or cells or in the form of cell lines, from animal or human origin, can be useful for the OOC approach. The basic criterion for selecting the stem cells for OOC is target disease. For the diseases with well‐known gene mutation, the DNA, specific disease‐ causing DNA mutation can be introduced into a stem cell line by the technique of homologous

cardiac tissue aligned with multiple hiPSC cardiac cells layer [103].

membrane‐based platform for stem cells growth.

*3.2.5. Stem cells on chip*

98 Lab-on-a-Chip Fabrication and Application

As described in previous section, microengineering platforms evolved as critical methods for the fabrication of various models of organs in the biomedical sciences. Newer inventions in this filed are reported to generate patterns of complex microstructures with precise control of fluid dynamics and incorporation of specific biological element that simulates organ functions directly. In this segment, few OOC devices based on anatomical mimicking will be described.

#### *3.3.1. Spleen on chip*

Spleen is a secondary lymphoid organ for selective filtration of damaged RBCs and infectious microbes including *Plasmodium* parasites [107]. Keeping in mind its special role in filtration and to understand its functionally in deeper sense, it was critical to design an OOC with high precision and accuracy. Spleen consists of white pulp, red pulp, and the marginal zone and slow blood microcirculation through the reticular meshwork of the splenic red pulp with increasing hematocrit is the prime reason of its unique filtering capacity that facilitates specialized macrophages in recognizing and destroying unhealthy RBCs [108]. Rigat‐Brugar‐ olas et al. designed a novel microdevice to copy the physical properties and hydrodynamic forces of the splenon; the minimal functional unit of the red pulp able to maintain filtering functions (**Figure 16a**) [108]. Their design consists of two main microfluidic channels for flow division to mimic the closed fast and the open slow microcirculations of spleen. The junction between slow‐flow and fast‐flow channel was arranged with parallel 2 μm microconstrictions resembling the IES to constrain cells. This device could precisely reproduce the natural physiochemical conditions of spleen and the unique characteristic of distinguishing different RBCs based on their mechanical properties.

**Figure 16.** Splenon on a chip: (a‐left) Diagram of the human splenon showing the closed‐fast and open‐slow microcir‐ culations as well as the interendothelial slits (IES); (a‐right) Schematic representation of flow division zone, the pillar matrix and microchannels within slow‐flow channel to mimic IES, respectively [108], (b) Artery on a chip: Schematic representation of a resistance artery segment on a chip contains a microchannel network, an artery loading well and an artery inspection area. ECs and SMCs represent the endothelial cells and smooth muscle cells, respectively [110].

#### *3.3.2. Blood vessel on chip*

*3.3.1. Spleen on chip*

100 Lab-on-a-Chip Fabrication and Application

RBCs based on their mechanical properties.

Spleen is a secondary lymphoid organ for selective filtration of damaged RBCs and infectious microbes including *Plasmodium* parasites [107]. Keeping in mind its special role in filtration and to understand its functionally in deeper sense, it was critical to design an OOC with high precision and accuracy. Spleen consists of white pulp, red pulp, and the marginal zone and slow blood microcirculation through the reticular meshwork of the splenic red pulp with increasing hematocrit is the prime reason of its unique filtering capacity that facilitates specialized macrophages in recognizing and destroying unhealthy RBCs [108]. Rigat‐Brugar‐ olas et al. designed a novel microdevice to copy the physical properties and hydrodynamic forces of the splenon; the minimal functional unit of the red pulp able to maintain filtering functions (**Figure 16a**) [108]. Their design consists of two main microfluidic channels for flow division to mimic the closed fast and the open slow microcirculations of spleen. The junction between slow‐flow and fast‐flow channel was arranged with parallel 2 μm microconstrictions resembling the IES to constrain cells. This device could precisely reproduce the natural physiochemical conditions of spleen and the unique characteristic of distinguishing different

**Figure 16.** Splenon on a chip: (a‐left) Diagram of the human splenon showing the closed‐fast and open‐slow microcir‐ culations as well as the interendothelial slits (IES); (a‐right) Schematic representation of flow division zone, the pillar matrix and microchannels within slow‐flow channel to mimic IES, respectively [108], (b) Artery on a chip: Schematic representation of a resistance artery segment on a chip contains a microchannel network, an artery loading well and an artery inspection area. ECs and SMCs represent the endothelial cells and smooth muscle cells, respectively [110].

Arteries or blood vessels transport the blood in human body. Geometry of vasculature and accumulation of particles inside the vessels varies with the pathological changes in the structure and function of small blood vessels, which leads to cardiovascular diseases [109]. Scalable approaches to assess the structure and function of intact cardiovascular tissues in health and disease will be crucial for developing better treatment strategies. Fluid sheer stress and cyclic stretch are other parameters that should be taken into account while designing *in vitro* vessels on‐chip systems.

Typically, most of the current systems contain small arteries mounted on two wires or perfused with glass micropipettes that suffer from the disadvantages of nonscalability and need of a skilled person to operate. To overcome this barrier, Gunther et al. presented a scalable organ‐ based microfluidic platform for loading, precise placement, fixation as well as controlled perfusion and superfusion of a fragile resistance artery segment (**Figure 16b**) [110]. This device was comprised of three parts: the artery‐loading area, a microchannel network and a separate artery inspection area, connected to a thermoelectric heater and a thermoresistor to maintain the temperature at 37°C. Resistance arteries had specialized structures with 30–300 μm diameters to regulate the flow and redistribution of blood in organs. As depicted, the setup was located in the terminal sections of the arterial vascular tree, and their walls are composed of a single layer of lining endothelial cells (ECs) and several layers of circumferentially arranged smooth muscle cells (SMCs). This device although could not replicate the full functionality but showed a unique property to analyze small artery structure and function through exposure to a well‐defined heterogeneous spatiotemporal microenvironment.

In another approach by Zhang et al., cyclic stretching of vesicular endothelial cells can be studied. They designed a two‐layered microsystem with upper microfluidics layer and bottom groove layer separated by an elastic membrane to provide cyclic stretch (**Figure 17**). A vacuum pump was integrated with the device to apply suction pressure on membrane resulting in cyclic stretch [111].

#### **3.4. Perfusion‐based on‐chip systems**

Cell‐cell interactions are vital for maintaining tissue structure and function, and many cells respond to both homotypic and heterotypic interactions. Combining fluid flow and mechanical forcing regimens as in *in vivo* cellular environment can improve tissue‐ and organ‐specific functions [66]. In this section, we describe few microengineering systems for liver, brain and womb that were designed for better understanding of mechanism of cellular interactions [82].

#### *3.4.1. Liver on chip*

Liver is considered to be one of the versatile organ performing thousands of functions that include detoxification, protein synthesis, hormone production, glycogen storage, etc. It is also a key player in human drug interaction and a trivial target for drug‐induced toxicity.

Liver possess a complex structure and hepatic lobule is its prime functional unit consisting of hepatocytes, blood vessels, sinusoids and Kupffer cells. [112]. Hepatocytes are crucial con‐

**Figure 17.** Schematic of blood vessel on chip. (a) PDMS chambers connected by a membrane, (b) Fabricated device, (c) Microfluidic channel for consecutive flow, (d) Stretching and relaxed elastic membrane [111].

tributors to liver functions and necessary for understanding the metabolism of xenobiotics and possible hepatotoxic effects in pharmacology. However, hepatocytes lack proliferative properties and biological interactions, which makes it rather difficult to maintain the liver‐ specific function of these cells *in vitro.* As a solution to this barrier, Kane et al. demonstrated a microfluidic array with wells capable of supporting micropatterned primary rat hepatocytes in coculture with 3T3‐J2 fibroblasts [113]. In this process, under continuous perfusion with medium and oxygen, the synthetic and metabolic capacity of hepatocytes were preserved as evidenced by the continuous and steady synthesis of albumin and production of urea.

In other approach by Du et al., encapsulated hepatocytes that were produced with recombinant protein, with endothelial cells, differentiating them from hiPSCs within specific niches in multicomponent hydrogel fibers and further assembled into 3D‐patterned endothelialized liver tissue constructs [114]. Endothelial cells significantly improved the function of hepato‐ cytes *in vitro* and when tested on a mouse model of partial hepatectomy, an improved vascularization of the fiber scaffold was observed.

A miniaturized, multiwall coculture system for human hepatocytes surrounded by fibroblasts with optimized microscale architecture that maintained the typical phenotypic functions of the hepatocytes for several weeks was reported by Bhatia et al. Another device comprised of three sections, including a central channel for heptocytes, a microfluidics convection channels and a microfluidics sinusoid barrier with a set of narrow channels to model epithelial cells as show in **Figure 18a** [115]. This model succeeded in mimicking the transportation between blood flow and hepatocytes and the sheer stress experienced by hepatocyts.

**Figure 18.** (a) Schematic of soft lithographic process to fabricate microscale multiwell format for primary hepatocytes that selectively adhere to matrix‐coated domains and coculture with fibroblasts seeded on bare areas [115]. (b) Config‐ uration of one basic unit of liver tissue, the classic hepatic lobule and lobule‐mimetic‐stellate‐electrodes array. (c) The configuration and operation principles of DEP‐based heterogeneous lobule‐mimetic cell patterning [116].

tributors to liver functions and necessary for understanding the metabolism of xenobiotics and possible hepatotoxic effects in pharmacology. However, hepatocytes lack proliferative properties and biological interactions, which makes it rather difficult to maintain the liver‐ specific function of these cells *in vitro.* As a solution to this barrier, Kane et al. demonstrated a microfluidic array with wells capable of supporting micropatterned primary rat hepatocytes in coculture with 3T3‐J2 fibroblasts [113]. In this process, under continuous perfusion with medium and oxygen, the synthetic and metabolic capacity of hepatocytes were preserved as evidenced by the continuous and steady synthesis of albumin and production of urea.

**Figure 17.** Schematic of blood vessel on chip. (a) PDMS chambers connected by a membrane, (b) Fabricated device, (c)

Microfluidic channel for consecutive flow, (d) Stretching and relaxed elastic membrane [111].

In other approach by Du et al., encapsulated hepatocytes that were produced with recombinant protein, with endothelial cells, differentiating them from hiPSCs within specific niches in multicomponent hydrogel fibers and further assembled into 3D‐patterned endothelialized liver tissue constructs [114]. Endothelial cells significantly improved the function of hepato‐ cytes *in vitro* and when tested on a mouse model of partial hepatectomy, an improved

A miniaturized, multiwall coculture system for human hepatocytes surrounded by fibroblasts with optimized microscale architecture that maintained the typical phenotypic functions of the hepatocytes for several weeks was reported by Bhatia et al. Another device comprised of three sections, including a central channel for heptocytes, a microfluidics convection channels and a microfluidics sinusoid barrier with a set of narrow channels to model epithelial cells as

vascularization of the fiber scaffold was observed.

102 Lab-on-a-Chip Fabrication and Application

Another research, also based on hepatocytes‐based model, was done by Ho et al., where they designed an array of concentric‐stellate‐tip microelectrodes to mimic the lobular structure of liver tissues (**Figure 18b**, **c**) [116]. This device was comprised of vertical microelectrodes or lobule‐mimicking stellate electrode arrays, to achieve 3D liver cell patterning by separately snaring hepatocytes and endothelial cells that were manipulated under patterned electric fields via dielectrophoresis (DEP). Few other researchers (i.e., Feng et al. [117], Wong et al. [118], Lee et al. [119]) have also put forth their proof of concepts based on hepatocytes. Wong et al. developed a concave microwell‐based size controllable spheroidal "hepatosphere" and "heterosphere" models by monoculturing primary hepatocytes and by coculturing primary hepatocytes and hepatic stellate cells (HSCs), respectively, to monitor the effect of HSCs in controlling the formation of tight cell‐cell contacts and final organization of the spheroidal aggregates [118, 82].

Some other reports are also there where researcher came forward with their ideas to design efficient liver on‐chip devices for drug screening and toxicity analysis [98, 120–122].

Recently, Lee et al. have designed a novel liver on‐chip system based on liver microsomes that were encapsulated in 3D hydrogel matrix to mimic the metabolism reactions and the transport phenomena in the liver. Photopolymerization of poly(ethylene glycol) diacrylate (PEG‐DA) allowed controlling the mass transfer with matrix sizes. To reproduce the blood flow through liver, gravity‐induced passive flow was explored. They measured the reaction kinetics of P450 enzymes in the device and simulated the convection‐diffusion‐reaction characteristics inside the device with a mathematical model [123]. **Figure 19a** is illustrating the schematic and design of on‐chip liver platform. Although there were several factors to be modified for improved reaction kinetic data such as diffusion limitation, optimization of convection and mixing, reducing the nonspecific binding to PDMS surface, preliminary analysis shows great potential and this device will be further explored for the metabolism of various compounds in liver [123].

**Figure 19.** (1) Schematic of PDMS chip fabrication method and picture of fabricated chip (size of the glass slide was 25 mm by 75 mm) [123]. (2) Schematic diagram illustrating the sequential procedure for constructing the biomimetic mi‐ crotissue [124].

Most of the on‐chip liver platforms are based on hepatocytes, and generally, these *in vitro* hepatocyte culture systems imitated the structure of the hepatic cord or can applied for studying specific aspects of toxicity. However, to imitate advanced liver architectures (i.e., hepatic sinusoids) that could preserve cell‐cell and cell‐ECM interactions, these existing devices did not solve the purpose. To overcome this limitation, Ma et al designed a microflui‐ dics‐based biomimetic method for *in vitro* fabrication of a 3D liver lobule such as microtissue. Their system was composed of a radially patterned hepatic cord‐like network and an intrinsic hepatic sinusoid‐like network as shown in **Figure 19b**. This device showed that the 3D biomimetic liver lobule‐like microtissue retained higher basal liver‐specific functions in Phase I/II (i.e., CYP‐1A1/2 and UGT activities) and more sensitive response was obtained for pharmacological inducers/inhibitors than the 2D and 3D monocultures of HepG2 cells. This device was tested for three model drugs—acetaminophen, isoniazid and rifampicin and a high hepatic capacity for drug metabolism was exhibited by biomimetic microtissue that indicated that microtissue, designed by Ma et al. can be explored as a promising platform for *in vitro* toxicity of drugs [124].

#### *3.4.2. Brain on chip*

Human brain is the most complex structure and the quest to understand how it stores and processes information leads researches to the application of new microengineering technolo‐ gies to design *in vitro* model of brain. Unraveling the basic concepts could be beneficial for neural diseases, development of improved brain‐machine interfaces and domain of machine‐ learning will be totally revolutionized. A brain‐oriented paradigm shift has occurred with the advances in neuroscience and OOC systems [131, 132].

enzymes in the device and simulated the convection‐diffusion‐reaction characteristics inside the device with a mathematical model [123]. **Figure 19a** is illustrating the schematic and design of on‐chip liver platform. Although there were several factors to be modified for improved reaction kinetic data such as diffusion limitation, optimization of convection and mixing, reducing the nonspecific binding to PDMS surface, preliminary analysis shows great potential and this device will be further explored for the metabolism of various compounds in liver [123].

**Figure 19.** (1) Schematic of PDMS chip fabrication method and picture of fabricated chip (size of the glass slide was 25 mm by 75 mm) [123]. (2) Schematic diagram illustrating the sequential procedure for constructing the biomimetic mi‐

Most of the on‐chip liver platforms are based on hepatocytes, and generally, these *in vitro* hepatocyte culture systems imitated the structure of the hepatic cord or can applied for studying specific aspects of toxicity. However, to imitate advanced liver architectures (i.e., hepatic sinusoids) that could preserve cell‐cell and cell‐ECM interactions, these existing devices did not solve the purpose. To overcome this limitation, Ma et al designed a microflui‐ dics‐based biomimetic method for *in vitro* fabrication of a 3D liver lobule such as microtissue. Their system was composed of a radially patterned hepatic cord‐like network and an intrinsic hepatic sinusoid‐like network as shown in **Figure 19b**. This device showed that the 3D biomimetic liver lobule‐like microtissue retained higher basal liver‐specific functions in Phase I/II (i.e., CYP‐1A1/2 and UGT activities) and more sensitive response was obtained for pharmacological inducers/inhibitors than the 2D and 3D monocultures of HepG2 cells. This device was tested for three model drugs—acetaminophen, isoniazid and rifampicin and a high hepatic capacity for drug metabolism was exhibited by biomimetic microtissue that indicated that microtissue, designed by Ma et al. can be explored as a promising platform for *in vitro*

Human brain is the most complex structure and the quest to understand how it stores and processes information leads researches to the application of new microengineering technolo‐ gies to design *in vitro* model of brain. Unraveling the basic concepts could be beneficial for

crotissue [124].

104 Lab-on-a-Chip Fabrication and Application

toxicity of drugs [124].

*3.4.2. Brain on chip*

**Figure 20.** (A) The microfluidic‐based culture platform directs axonal growth of CNS neurons and fluidically isolates axons [125]. (B) Schematic diagrams of normal brain mimicking microfluidic chip (a) and Alzheimer's disease brain mimicking microfluidic chip [127].

We discussed earlier various other OOC but owing to its structural and functional hierarchy, high specialization and constant metabolic demand to design a complete *in vitro* brain model is difficult. The prime limiting factors are used to identify the smallest structural and functional unit, ion channels or synapses in the microenvironment [126]. Researchers from all over the world give different experimental models of circular microfluidic compartmentalized cocul‐ ture platforms to study brain development and degeneration based on physiological neuron connection architecture. A microfluidic culture platform was demonstrated by Taylor et al, consist of a relief pattern of somal and axonal compartments connected by microgrooves that function in directing, isolation and biochemical analysis of CNS axons (**Figure 20a**) [125]. In another work, Park et al describe a microfluidic chip based on 3D neurospheroids that more closely mimics the *in vivo* brain microenvironment and provides a constant flow of fluid similar in the interstitial space of the brain. Concave microwell arrays were explored for the formation of uniform neurospheroids, with cell‐cell interactions and contacts in all directions while osmotic micropump was used to maintain the slow interstitial level of flow. Using this platform, effect of flow on neurospheroid size, neural network and neural differentiation was investigated via this *in vitro* platform. Larger sizes of neurospheroids were obtained and formed more robust and complex neural networks than those cultured under static conditions. This finding proved the effect of the interstitial level of slow and diffusion‐dominant flow on continuous nutrient, oxygen and cytokine transport and removal of metabolic wastes [127]. This chip was designed to detect the toxic effect of β‐amyloid; a major contributor of Alz‐ heimer's disease. **Figure 20b** is showing the schematic of this unique platform for neurodege‐ nerative disease diagnostic. Kato‐Negishi et al. came up with a millimeter‐sized neural building block to reconstruct 3D broad neural networks connecting with different neurons [128]. Peyrin et al. also described a microfluidic system involving several different neuron subtypes separated into two individual chambers with asymmetrical connection architecture of funnel‐shaped microchannels to reconstruct oriented neuronal networks [129]. This device was a kind of diode that operated as direction selective filter where axonal projections can be penetrated by axons in a single direction and as an impermeable barrier for cell bodies. In this point, Kunze et al. demonstrated a 3D microfluidic device for creating physiologically realistic, micrometer scaled neural cell multilayers in an alginate‐enriched agarose scaffold [82, 130].

A method to fabricate neurospheres networking with nerve‐like structure using concave well arrays connected by the hemicylindrical channels was illustrated by Jeong et al. This method provides the topological effect of the concave‐well hemicylindrical‐channel‐networking, which is crucial in guided outgrowth of neuronal network [131]. Similar hemicylindrical systems were also explored to generate 3D nerve‐like neural bundles between neural spheroids and neighboring satellite spheroids in concave channels [132].

#### *3.4.3. Breast and womb on chip*

Breast cancer is still the cause of concern and with the advancement of microfabrication techniques, improved detection and therapy of breast neoplasia can be obtained via nanode‐ vices traveling inside mammary ducts. However, the decreasing size of branched mammary ducts prevents access to remote areas of the ductal system using a pressure‐driven fluid‐based approach. Magnetic field guidance of superparamagnetic submicron particles (SMPs) in a stationary fluid might provide a possible alternative but it is critical to first reproduce the breast ductal system to assess the use of such devices for future therapeutic and diagnostic ("thera‐ nostic") purposes. Graften et al. came up with an idea of to engineer a portion of a breast ductal system using polydimethylsiloxane (PDMS) microfluidic channels of decreasing sizes with a total volume of 0.09 mL. A magnet was used to move superparamagnetic/fluorescent SMPs through a static fluid inside the microchannels [133]. **Figure 21a** is the schematic of PDMs on chip assembly. This device can be explored for the early detection of ductal breast cancer and consisted of basoapically polarized monolayer of luminal cells only as the device imitated the luminal portion of the ductal breast system only and myoepithelial cells at the basal side of the luminal cells and terminal ductal lobular units at the ends of the narrowest channels were not included. Apart from breast on chip, womb OOC was also developed by Chang et al. with the objective to deal with infertility.

In recent years, a genuine increase in infertility has been observed due to diverse factors, including stress, environmental pollution and increase in age, smoking, consumption of alcohol, sexually transmitted diseases, etc. *In vitro* fertilization (IVF), a state‐of‐the‐art tech‐ nology, enhances the rate of pregnancy. As a procedure, fertilized eggs in the blastocyst stage are transferred to the woman's uterus for implantation and further development and efforts are made to improve the culture environment of the preimplantation embryos and developing specialized culture surfaces to enhance the success rate of this technique [134, 135].

Due to the failure of static culture systems to mimic the dynamic fluid environment in the fallopian tube [136], dynamic culture platforms that explored shaking/rotation [137], control‐ led fluid flow [138] and vibration [139] models were studied for use in embryonic development where method of coculturing embryos with endometrial was done to overcome developmental arrest of early embryos in single culture. Although these methods showed enhanced perform‐

**Figure 21.** (a) Schematic of breast on chip [133], (b) PDMS‐based embryo coculture microchip, where the concentration gradient generator is integrated with a mixer and a cell culture chamber on the top [140].

ance and beneficial effects of coculturing on the development of mammalian, they could not be considered as a complete on chip system for womb. Recently, Chang et al. also designed an autologous 3D perfusion platform as a necessary approach to deal with IVF and partly mimic the physiological function of the reproductive system [140]. This device as shown in **Fig‐ ure 21b** is comprised of an upstream concentration gradient generator (width: 250 μm, height: 230 μm) was integrated with a diamond‐shaped passive micromixer (width: 200 μm, height: 230 μm) that could generate six different homogeneous concentrations of progesterone. Micromixer was used to increase the contact area between liquid molecules and to provide enhanced mixing efficiency by its continuous splitting and mixing of liquids. The main specifications and goals of this microfluidic channel design was as follows: (i) Gradient distribution for specific concentrations of steroid hormones in six culture chambers, (ii) Maintaining homogenous concentrations of steroid hormones in individual chamber, (iii) Preserve uniform culture conditions with respect to the flow speed/rate by constant flow speed/rate for the chambers.

This womb‐on‐chip platform showed the ability to replace the present embryo culture platforms used for assisting *in vitro* fertilization.

#### **3.5. Human on chip**

of funnel‐shaped microchannels to reconstruct oriented neuronal networks [129]. This device was a kind of diode that operated as direction selective filter where axonal projections can be penetrated by axons in a single direction and as an impermeable barrier for cell bodies. In this point, Kunze et al. demonstrated a 3D microfluidic device for creating physiologically realistic, micrometer scaled neural cell multilayers in an alginate‐enriched agarose scaffold [82, 130]. A method to fabricate neurospheres networking with nerve‐like structure using concave well arrays connected by the hemicylindrical channels was illustrated by Jeong et al. This method provides the topological effect of the concave‐well hemicylindrical‐channel‐networking, which is crucial in guided outgrowth of neuronal network [131]. Similar hemicylindrical systems were also explored to generate 3D nerve‐like neural bundles between neural spheroids

Breast cancer is still the cause of concern and with the advancement of microfabrication techniques, improved detection and therapy of breast neoplasia can be obtained via nanode‐ vices traveling inside mammary ducts. However, the decreasing size of branched mammary ducts prevents access to remote areas of the ductal system using a pressure‐driven fluid‐based approach. Magnetic field guidance of superparamagnetic submicron particles (SMPs) in a stationary fluid might provide a possible alternative but it is critical to first reproduce the breast ductal system to assess the use of such devices for future therapeutic and diagnostic ("thera‐ nostic") purposes. Graften et al. came up with an idea of to engineer a portion of a breast ductal system using polydimethylsiloxane (PDMS) microfluidic channels of decreasing sizes with a total volume of 0.09 mL. A magnet was used to move superparamagnetic/fluorescent SMPs through a static fluid inside the microchannels [133]. **Figure 21a** is the schematic of PDMs on chip assembly. This device can be explored for the early detection of ductal breast cancer and consisted of basoapically polarized monolayer of luminal cells only as the device imitated the luminal portion of the ductal breast system only and myoepithelial cells at the basal side of the luminal cells and terminal ductal lobular units at the ends of the narrowest channels were not included. Apart from breast on chip, womb OOC was also developed by Chang et al. with

In recent years, a genuine increase in infertility has been observed due to diverse factors, including stress, environmental pollution and increase in age, smoking, consumption of alcohol, sexually transmitted diseases, etc. *In vitro* fertilization (IVF), a state‐of‐the‐art tech‐ nology, enhances the rate of pregnancy. As a procedure, fertilized eggs in the blastocyst stage are transferred to the woman's uterus for implantation and further development and efforts are made to improve the culture environment of the preimplantation embryos and developing

Due to the failure of static culture systems to mimic the dynamic fluid environment in the fallopian tube [136], dynamic culture platforms that explored shaking/rotation [137], control‐ led fluid flow [138] and vibration [139] models were studied for use in embryonic development where method of coculturing embryos with endometrial was done to overcome developmental arrest of early embryos in single culture. Although these methods showed enhanced perform‐

specialized culture surfaces to enhance the success rate of this technique [134, 135].

and neighboring satellite spheroids in concave channels [132].

*3.4.3. Breast and womb on chip*

106 Lab-on-a-Chip Fabrication and Application

the objective to deal with infertility.

Organ‐on‐chip concept is in its nascent state and despite of the substantial advances in the creation of microengineered tissue and organ models, a lot is left to explore for recreating complex 3D models that could reconstitute the whole organ metabolism and physiology. With the recent advances in tissue engineering, microfabrication techniques, researchers are now focusing on multiorgan‐on‐chip devices that could imitate complete human on chip up to some extent if not fully [141, 142]. Figure shows a body‐on‐chip systems.

Although complete functional body‐on‐chip devices are still far from reach but the latest development in this field has given a glimpse of promising future of this revolutionary field of biomedicine. **Figure 22** shows the concept of body‐on‐chip microsystem [4].

**Figure 22.** Schematic of a body on chip system. (a) A microdevice containing interconnected cell culture microcham‐ bers integrated with microfluidic culture of intestinal epithelial, hepatocytes and breast cancer cells. (b) A micro cell culture analog (μCCA) representing a colon tumor, the bone marrow and liver [4].
