Current and Future Cornea Chip Models for Advancing Ophthalmic Research and Therapeutics

Abstract

Corneal blindness remains a significant global health challenge, with limited treatment options due to donor tissue scarcity outside of the United States and inadequate in vitro models. This review analyzes the current state of cornea chip technology, addressing fundamental challenges and exploring future directions. Recent advancements in biomaterials and fabrication techniques are discussed that aim to recapitulate the complex structure and function of the human cornea, including the multilayered epithelium, organized stroma, and functional endothelium. The review highlights the potential of the cornea chips to revolutionize ocular research by offering more predictive and physiologically relevant models for drug screening, disease modeling, and personalized medicine. Current designs, their applications in studying drug permeability, barrier function, and wound healing, and their limitations in replicating native corneal architecture, are examined. Key challenges include integrating corneal curvature, basement membrane formation, and innervation. Applications are explored in modeling diseases like keratitis, dry eye disease, keratoconus, and Fuchs' endothelial dystrophy. Future directions include incorporating corneal curvature using hydraulically controlled systems, using patient‐derived cells, and developing comprehensive disease models to accelerate therapy development and reduce reliance on animal testing.

This review analyzes cornea chip technology as an innovative solution to corneal blindness and tissue scarcity. The examination encompasses recent developments in biomaterial design and fabrication methods replicating corneal architecture, highlighting applications in drug screening and disease modeling while addressing key challenges in mimicking native corneal properties.

1. Introduction

Corneal blindness represents a significant global health challenge that affects millions of people worldwide and causing substantial socioeconomic burdens. Various diseases and injuries, including bacterial and fungal infections, corneal ectasia, dystrophies, and trauma, contribute to this pervasive issue.^{[ 1 ]} The World Health Organization estimates that corneal opacities account for 5.1% of global blindness, with a disproportionate impact on low‐ and middle‐income countries.^{[ 2 ]} Despite the pressing demand for novel therapies, the development of effective treatments still needs to be improved by several factors. The shortage of donor tissues for corneal transplantation outside of the United States limits the reach of this intervention.^{[ 3 ]} Furthermore, the lack of suitable in vitro models that accurately recapitulate human corneal complexity impedes both basic research and drug development efforts.

Animal models, including rabbits, mice, and pigs, have traditionally been employed in corneal research to study disease mechanisms, test drug efficacy, and develop new therapies.^{[ 4 ]} However, these models present several limitations that compromise their translational value. Significant species differences in corneal anatomy, physiology, and immune responses often lead to discrepancies between animal studies and human clinical trials.^{[ 5 ]} Furthermore, the use of animal models raises ethical concerns and is subject to increasing regulatory restrictions. For instance, the Food and Drug Administration Modernization Act 2.0 encourages the use of alternative testing methods to reduce reliance on animal studies.^{[ 6 ]}

Conventional in vitro models, including 2D cell cultures and ex vivo tissue explants, have also been utilized to study corneal biology and assess drug permeability.^{[ 7 ]} While these approaches provide valuable insights into cellular behavior and gene expression, they fall short in replicating the complexity and dynamic environment of the native cornea. 2D cell cultures fail to replicate the complex cell–cell and cell–matrix interactions essential for corneal homeostasis and wound healing.^{[ 8 ]} While ex vivo tissue explants offer greater physiological relevance, they suffer from limited viability and high intersample variability, making them unsuitable for both long‐term studies and high‐throughput screening.

Organ chip technology, particularly cornea chip models, has emerged as a promising alternative for studying corneal physiology, pathology, and drug screening.^{[ 9 ]} These platforms bridge the gap between traditional in vitro models and animal studies by replicating key aspects of the complex structure and functions of the human cornea in a controlled environment. Current cornea chips enable researchers to investigate drug transport, barrier integrity, and disease mechanisms with improved physiological relevance.^{[ 10 ]} However, further advancements are necessary to develop genuinely high‐fidelity cornea tissue models. Challenges lie in recapitulating the detailed extracellular matrix (ECM) organization and cellular functions. The corneal ECM, with its unique composition and arrangement of collagen fibrils, proteoglycans, and glycoproteins, plays a crucial role in maintaining corneal transparency, biomechanical properties, and cellular behavior.^{[ 11 ]} Additionally, the complex interactions between corneal epithelial, stromal, and endothelial cells, as well as their surrounding microenvironment, are essential for maintaining corneal homeostasis and wound healing.^{[ 12 ]}

This review critically analyzes the current state of cornea chip development, addressing fundamental challenges and exploring future directions in this rapidly evolving field. We will discuss how cornea chips can revolutionize ocular research by offering more predictive and physiologically relevant models for drug screening, disease modeling, and personalized medicine approaches. The review will cover recent advancements in biomaterials and fabrication techniques, the integration of corneal innervation and immune components, and the development of cornea chip systems. By highlighting the most promising developments and prospects, we aim to provide researchers and clinicians with a thorough understanding of the potential of cornea chip technology to accelerate the development of novel therapies for corneal diseases and reduce reliance on animal models in preclinical ocular medicine testing.

1.1. Essence of Cornea Anatomy

The cornea, a highly specialized transparent tissue forming the anterior portion of the eye, serves two critical functions: it maintains visual acuity by refracting light onto the retina and protects the eye from external factors, including dust, debris, and pathogens.^{[ 12 ]} The cornea comprises three main cellular layers: corneal epithelium, stroma, and corneal endothelium, as shown in Figure 1A. These layers are separated by two acellular membranes: the Bowman's layer (BL), which lies between the epithelium and stroma, and the Descemet's membrane (DM), which is situated between the stroma and endothelium.^{[ 13 ]} Each of these components has unique functions and structures that contribute to the overall health and transparency of the cornea. Overview of structural and cellular properties are summarized in Table 1 . The epithelium, the outermost layer, consists of a stratified squamous epithelium that acts as a barrier against infections and injuries.^{[ 12} , ^{14 ]} This protective barrier comprises three distinct sublayers: the basal, wing, and superficial layers, each with unique morphology and cell density.^{[ 15 ]} To maintain the health barrier, the epithelium is constantly renewed by limbal stem cells located at the corneal periphery, which maintain the epithelial cell population and prevent conjunctival invasion. The basal layer originates from limbal stem cells and serves as the progenitor source for both the wing and superficial cells, maintaining the regenerative capacity of the epithelium.^{[ 16 ]} The stroma comprises approximately 90% of the corneal thickness, consisting of highly organized collagen fibrils arranged in orthogonal lamellae to provide structural integrity and transparency.^{[ 11 ]} The collagen architecture of the stroma exhibits a gradient in density and organization. The anterior region features denser, more compact collagen fibrils, while the posterior region is looser with a more orthogonal alignment.^{[ 17 ]} Keratocytes, the primary cell type in the stroma, are interwoven within this collagen architecture and are responsible for maintaining ECM and responding to injuries through a complex wound‐healing process.^{[ 18 ]} The cell density of keratocytes decreases toward the posterior region of the stroma, reflecting its structural composition. The keratocytes can be transformed into fibroblasts and myofibroblasts, especially during wound healing, by various growth factors and cytokines secreted from the epithelium.^{[ 19 ]} The innermost layer, the endothelium, is a single layer of hexagonal morphology that lines the posterior surface of the cornea and maintains corneal hydration and clarity by regulating fluid balance through a combination of active ion transport and passive barrier functions.^{[ 20 ]} Cell density critically affects the corneal endothelium morphology due to tight and gap junctions. Unlike corneal epithelial and stromal cells, corneal endothelial cells do not have proliferative and regenerative functions, leading to decreased cell density with aging.^{[ 16 ]} Additionally, the peripheral region of the corneal endothelium has a denser structure than the central area.^{[ 21 ]} These native structural properties of the endothelium are crucial to corneal function.

Figure 1.

Recapitulation of human cornea anatomy and physiology. A) The structure of the mimicked in‐vitro corneal model. Schematic of the corneal model includes three vertically stacked compartments with three different cell types. B) Exploded view and schematics of cornea chip (B, right) with three microchannel compartments separated by hybrid interfacial membranes.

Table 1.

Corneal structure and cellular properties.

Layer		Thickness		Mechanical property	Components	Regeneration
Epithelium	Superficial	50 µm, 5–7 layers[ 45 ]	2–6 µm, 2–3 layers[ 45 ]	Elastic modulus: 0.57 kPa[ 103 ]	1227± 96.6 cells mm−2[ 104 ]	O
	Wing		2–3 layers[ 45 ]		4759± 380.5 cells mm−2[ 104 ]
	Basal		20 µm, single layer[ 45 ]		10479± 833 cells mm−2[ 104 ]
Basement membrane		40–60nm[ 45 ]		7.5± 4.2kPa[ 105 ]	Collagen type IV, laminin, perlecan and nidogen[ 45 ]	O
Bowman's layer		8–12µm[ 106 ]		109.8± 13.2kPa[ 107 ]	Collagen type I, III, V, VI[ 108 ]	X
Stroma	Anterior	450–500 µm[ 109 ]		281± 214kPa[ 110 ]	990± 233 cells mm−2[ 15 ]	O
	Posterior			89.5± 46.1 kPa[ 110 ]	554± 132 cells mm−2[ 15 ]
Descemet's membrane		3‐10 µm[ 45 ]		50± 17.8 kPa[ 110 ]	Collagen type IV, VIII, XII, laminin, perlecan, nidogens, vitronectin, fibronectin[ 111 ]	△
Endothelium		5 µm, single layer[ 45 ]		–	2500–3000 cells/mm2[ 15 ]	X
Nerves	Intraepithelial nerve fiber	–		–	–	O
	Sub‐basal nerve plexus	0.1–0.5 µm[ 112 ]		–	68.076 ± 2.327 no. mm−2[ 113 ]
	Sub‐epithelial nerve plexus	–		–	–
	Mid‐stromal nerve plexus	–		–	–
	Stromal nerve bundle	Fiber size: 0.5–5 µm[ 114 ]		–	–

1.2. Current Cornea Chip Models

Researchers have developed several in vitro modeling approaches to replicate the unique physiological and biological characteristics of the cornea, including 3D hydrogels, bioprinting,^{[ 22 ]} organotypic cultures,^{[ 23 ]} and self‐assembled ECM structures.^{[ 24 ]} These approaches have advanced to create models that incorporate all three cell types – epithelial, stromal, and endothelial cells – enabling the study of cell interactions and drug permeability.^{[ 25 ]} Early organotypic models used collagen gels and chondroitin sulfate matrices to mimic the 3D corneal structure.^{[ 26 ]} These models successfully replicated partial cornea structure and demonstrated inflammatory responses similar to human cornea. However, to fully replicate in vivo structural and mechanical characteristics, these models need to be improved in ECM organization and enhanced mechanical strength.^{[ 27 ]} Recent studies have explored diverse biomaterials^{[ 28 ]} specifically tailored to meet the unique requirements of the cornea for mechanical properties and transparency.^{[ 29 ]} Even though these novel biomaterials and tissue‐engineered approaches mimic cornea ECM components and structure, limitations persist in achieving well‐organized stromal layers, replicating stromal cell density gradients, and developing fully functional endothelium and epithelium. Thus, in this section, we will discuss how to create more advanced culture platforms that better consider the structural and cellular properties of the cornea, which is crucial for developing physiologically relevant models for drug testing and disease modeling.

Recently, significant progress has been made in developing cornea chips that closely mimic the anatomical and physiological properties of the human cornea, as shown in Tables 2 and  3 . These chip models have been designed to address specific research questions like drug permeability, barrier function, and disease modeling. While each model has its strengths and limitations, they collectively demonstrate the potential of cornea chips to advance our understanding of corneal biology and facilitate the development of new therapies.^{[ 30 ]} Several studies have successfully incorporated air‐liquid interface (ALI) conditions into cornea chip designs, enabling the investigation of drug transport and barrier integrity. In an ALI culture system, the apical surface of epithelial cells is exposed to air while maintaining nutrient supply from the basolateral side, promoting epithelial differentiation and stratification to form a mature, functional barrier.^{[ 31 ]} For example, Bennet et al. developed a micro‐engineered human corneal epithelium‐on‐a‐chip using immortalized human corneal epithelial cells cultured on two‐chambered organ chips separated by a porous polycarbonate membrane, shown in Figure 2A.^{[ 32 ]} The fibronectin‐coated polycarbonate membrane in Table 3 provided mechanical properties closely resembling the in vivo Bowman's layer (108 ± 11.4 kPa) and the anterior basement membrane (7.5 ± 4.2 kPa). This chip model effectively mimics the in vivo human corneal epithelium, utilizing the additive manufacturing concepts to form 5 layers of corneal epithelial cells. This chip is utilized for assessing the mass transport of two types of eye drops, highlighting its potential for drug permeability studies under various tear flow conditions to mimic drug clearance. The permeability patterns of hydrophilic (Zaditor) and lipophilic (Pred Forte) drugs in this chip aligned with the transport mechanisms observed in the human cornea. Although this work has valuable information about mass transport depending on tear clearance flow regimes, the absence of stromal and endothelial layers restricts its applicability for studying full corneal drug penetration or diseases affecting deeper corneal layers. Similarly, Seo et al. (Figure 2B) developed a human cornea chip model incorporating a stratified epithelium supported by a keratocyte‐laden matrix and simulated eye blinking motion.^{[ 9 ]} The blinking motion generated mechanical forces and pressure gradients (0.3–7 kPa) consistent with physiological conditions observed in vivo. Tear film formation, measured using optical coherence tomography, revealed an average thickness of 6 µm, closely matching the physiological tear film thickness in the human eye (5–10 µm). This model successfully recreated dry eye disease (DED) conditions, demonstrating osmolarity levels (351 mOsm L⁻¹) consistent with clinical data from DED patients. The model also exhibited strong physiological relevance in evaluating key pathophysiological features, including inflammatory responses and fluorescein staining patterns similar to those observed in clinical DED cases, as well as enabling the evaluation of lubricin as a potential therapeutic agent. Despite these significant advances toward physiologically relevant cornea chips, some limitations remain. The synthetic scaffold in the stromal layer used in this model shows limited replication of the native stromal gradient and ECM organization of the human cornea. Moreover, the absence of endothelium represents a critical limitation, as this layer plays a crucial role in regulating corneal hydration and maintaining transparency. Without this layer, this cornea chip cannot replicate hydration regulation and tear film stability for understanding hyperosmolarity and epithelial damage due to DED.

Table 2.

Current review of cornea chips.

		Chip type	Membrane type	Cell type	Advantages	Limitations
Cornea chip	Epithelium	Two chamber chip[ 32 ]	Polycarbonate (PC)	Immortalized human corneal epithelial cells (hTCEpi)	TEER measurement port Mimics tear flow Drug permeability Multilayered epithelium	Limited to epithelial studies Use a cell line 2D cell culture Planar surface
		Two chamber chip[ 52 , 115 ]	Polyethylene Terephthalate (PET)	Immortalized human cornea epithelial cells (HCE‐T)	Dynamic flow stimuli Drug permeability Analysis of Extracellular metabolite	Air‐liquid interface Use a cell line 2D cell culture Planar surface
	Stroma	Open‐top array curvature chip[ 51b ]	–	Primary human keratocytes (HK)	Curvature control Array chip Cornea ectasia modeling Cell phenotype transformation Orthogonal cell orientation ECM deformation	Short‐term culture 2D cell culture
	Co‐culture	Open‐top chip with scaffold and blinking systems[ 9 ]	Polystyrene cell culture substrates	Primary human corneal epithelial cells Primary Human Keratocyte (HK)	Recreates tear film dynamics Dry eye disease modeling Mimicking blinking mechanism Air‐liquid interface Tear film formation Curvature‐shaped scaffold	Complex setup Randomized stromal cell orientation
		Open‐top chip[ 33 ]	Polycarbonate (PC)	Immortalized human corneal epithelial cells Immortalized Human corneal endothelial cells	Air–liquid interface TEER measurement port Scratch wound healing model The use of Extracellular vesicles for wound healing effectiveness	Missing stromal layer, including wound healing pathway 2D cell culture Planar surface Use a cell line
		Horizontally organized three chamber chip[ 36 ]	Viscous finger patterned channel	Primary mouse corneal epithelial cells Primary mouse corneal endothelial cells	Replicates full corneal thickness Allows for cell–cell interaction studies Drug permeability testing Easy cell imaging	Use murine cells, not human 2D cell culture Planar surface Simplified cornea stromal layer with collagen mixtures Nonuse of corneal stromal cells
Tissue engineering	Co‐culture	Transwell[ 43 ]	Polycarbonate (PC)	Primary human corneal Fibroblasts (hCFs) Immortalized human corneal epithelial cells (hCE‐TJ)	Co‐culture of hCFs and hCE‐TJ Long‐term culture to create self‐assemble cornea stroma Air–liquid interface Enhance cell‐to‐cell interaction between corneal epithelium and stroma Membrane formation	Absence of corneal endothelial layer. Limited multicellular stratification Planar surface
		Silk sponge and porous patterned silk films[ 44 ]	Stamped porous silk film	Human corneal stromal stem cells (hCSSCs) Human corneal epithelial cells (hCECs) Chicken dorsal root ganglion cell culture (DRG neurons)	Co‐culture of hCSSCs, hCECs, and DRG neurons Use of silk biomaterial similar to mechanical properties of human cornea Stamped silk films with varying nerve growth factor concentrations effectively guide neuronal innervation. Enhanced air–liquid interface with integrated innervation.	Use chicken DRG neurons Innervation density and length not quantified Absence of corneal endothelial layer. Planar surface
	Full cornea	Plastic dishes[ 25a ]	Nitrocellulose membrane	Primary bovine corneal epithelial cells Primary bovine corneal keratocytes Primary bovine corneal endothelial cells	Fully reconstruction of cornea in vitro Stromal cells embedded 3D collagen layer Air–liquid interface	Use bovine cells Planar surface Randomly organized cornea stromal layer Drug permeability testing not possible
		Collagen‐chondroitin sulfate substrate[ 26 ]		Immortalized Human cornea epithelial cells Immortalized human cornea keratocytes Immortalized human cornea endothelial cells	Fully reconstruction of cornea in vitro Air–liquid interface Comparison of corneal equivalent with Human cornea Inflammatory response with the human cornea by injury. Light transmission with the human cornea	Use a cell line Non‐self‐assembly Each cell is embedded in a collagen and chondroitin sulfate substrate

Table 3.

Mechanical and optical properties of interfacial membranes in various cornea chips.

Membrane type	Pore size [µm]	Thickness [µm]	Elastic modulus [Pa]	Water content [%]	Transparency [%]
Polycarbonate (PC)	5[ 32 ]	10	108 k[ 32 ]	–	–
Polyethylene Terephthalate (PET)	0.4[ 52 ]	10[ 52 ]	–	–	–
Polystyrene (Alvetex)	36–40	200	77 k[ 116 ]	–	85–90[ 117 ]
Silk Fibroin films	–	50–70[ 118 ]	67.7 k[ 44 ]	14.2–57.6[ 118 ]	90[ 118 ]
Nitrocellulose (Whatman)	0.45[ 119 ]	125[ 119 ]	12.2 M – 1.1G[ 119 ]	–	–

Figure 2.

The representative current cornea chip designs. A) Schematic of a two‐chamber microfluidic device with a polycarbonate membrane, designed to control tear flow. Reproduced with permission.^{[ 32 ]} Copyright 2018, Royal Society of Chemistry. B) The blinking ocular surface model incorporates a dome‐shaped 3D polystyrene scaffold with cells embedded in hydrogel using a compression approach. Reproduced with premission^{[ 9 ]} Copyright 2019, Springer Nature. C) The open‐top design provides an air‐liquid interface for corneal epithelial cells and includes a TEER measurement port. Reproduced with permission.^{[ 33 ]} Copyright 2022, Elsevier. D) Horizontally organized microfluidic chambers on a chip, fabricated using diverse soft‐lithography techniques, enable the co‐culture of three cell types and facilitate cellular investigation. Reproduced under terms of the CC‐BY license.^{[ 36 ]} Copyright 2020, Royal Society of Chemistry.

Several studies have developed co‐culture models to recapitulate the complex interactions within the cornea, expanding beyond corneal epithelium‐focused models. These co‐cultured chips incorporate epithelial, stromal, and neuronal cells, providing insights into corneal homeostasis and wound healing mechanisms. Yu et al. developed a comprehensive human cornea chip that incorporated both epithelial and endothelial layers (Figure 2C), as well as a collagen‐coated porous membrane to mimic an interfacial membrane.^{[ 33 ]} The chip featured compartmentalized channels for co‐culture of different cell types and an open‐top structure to provide ALI for the corneal epithelial cells. This model demonstrated the formation of a stratified epithelium, expression of cornea‐specific markers, and the establishment of a functional barrier. The barrier function of this cocultured model achieved trans‐epithelial electrical resistance (TEER) values of about 862 Ω cm², and the permeability coefficient of the model was 6.63 x 10⁻⁷ cm s ⁻¹. This chip was used to study the impact of extracellular vesicles (EV) derived from mesenchymal stem cells for corneal wound healing, showing that the EV could promote epithelial wound closure and reduce inflammation. Although the permeability coefficient of this cocultured chip is similar to that of the ex vivo human cornea model using sodium fluorescein molecule tracer (6.0 x 10⁻⁷ cm s⁻¹),^{[ 34 ]} the corresponding in vivo permeability coefficient has not yet been clearly defined. Additionally, while in vitro TEER values for the epithelium are widely recognized for their strong physiological relevance to in vivo conditions, the measured TEER values of the co‐culture model fall outside the reported in vivo range of 690 ± 69 Ω cm².^{[ 35 ]} This discrepancy reflects the absence of a stromal layer with proper ECM structure, which may limit its physiological relevance for simulating wound healing processes. Similarly, Bai et al. introduced a novel cornea chip design (Figure 2D) replicating 70–80% of the human corneal stroma layer by optimizing channel distance for drug permeability testing.^{[ 36 ]} Each channel of the chip was filled with isolated corneal epithelial cells from mice, a stromal layer composed of collagen hydrogel without cells, and isolated corneal endothelial cells within a horizontally designed three‐chambered structure. The chip design features microfluidic pores between each channel, mimicking BL by collagen coating through a viscous finger patterning method.^{[ 37 ]} This study is notable for its ability to simulate drug diffusion across the corneal barrier, demonstrating size‐dependent permeability for dextran molecules of 10, 40, and 70 kDa. It offers a distinct advantage over vertically stacked chips by enabling better microscopic investigation of cell–cell interactions at the interface. While the model verified the importance of epithelium in full cornea permeability through dextran molecule testing, the lack of detailed physiological relevance data remains a limitation. Furthermore, the use of murine cells instead of human cells may limit its translational potential, and the flat design fails to mimic the mechanobiological complexity of the cornea, particularly by neglecting the gravitational effects that influence in vivo cell organization and tissue architecture.

These advances in cornea chip technology have enabled valuable insights into drug transport and cellular behavior. However, the accurate replication of corneal ECM remains a significant challenge. Unlike the epithelium layer, the stromal layer in these current chips is often substituted with collagen hydrogels or human corneal stromal cells embedded within collagen‐based hydrogels. To further enhance the biomimetic structure of these chips, tissue engineering approaches offer promising strategies to recapitulate the ECM composition and organization of the native corneal stroma.^{[ 24 ]} Zhang and Chen et al. developed a corneal stromal model using patterned silk fibroin membranes.^{[ 38 ]} The dome‐shaped and stacked silk fibroin membranes exhibited native‐like collagen fibril organization and mechanical properties, providing a suitable environment for corneal keratocytes. The silk fibroin membrane achieved over 90% light transmission, comparable to ~90% transmission observed in human corneal tissue within the visible spectrum.^{[ 39 ]} Additionally, Gouveia et al. created a high‐density collagen gel by compressing it and modulating its stiffness through collagenase treatment to deform the mechanical properties intentionally.^{[ 40 ]} The compressed collagen in the central area of this model exhibited 4.8 ± 3.5 MPa, and collagenase‐treated gel demonstrated 0.7 ± 0.4 MPa, successfully recapitulating the relative mechanical differences of BL structure. However, the use of a homogeneous collagen hydrogel does not fully capture the complex, heterogeneous nature of the native corneal stroma. To address this, self‐assembling corneal stromal layers without synthetic materials have emerged as a promising approach for mimicking the native ECM structure.^{[ 41 ]} This approach promotes collagen fibril organization, keratocyte alignment, and cell‐to‐ECM interaction. Mechanical stimuli, including regulated compression and tension forces, mimic the physiological stresses experienced by the cornea in vivo. Miotto et al. demonstrated a self‐assembled structure (Figure 3A) by modulating a shape similar to corneal curvature using the Arg‐Gly‐Asp‐Ser peptide, which facilitates cell adhesion.^{[ 42 ]} The compressed collagen hydrogel with the peptide demonstrated significantly less contraction than collagen‐only hydrogel by interacting with corneal stromal cells. By modulating contraction in the central and peripheral areas, they formed a self‐assembled corneal stroma, where mechanical stimulations guided the orientation of collagen fibers and cellular alignment, crucial for replicating the natural structure and function of the corneal stroma.

Figure 3.

Adaptable Tissue‐engineered techniques on Cornea chip model. A) Self‐assembled curvature, regulated by stiffness through cell properties and RGD peptide interaction. Reproduced with permission.^{[ 42 ]} Copyright 2019, Wiley. B) The 3D in vitro cornea model, incorporating natural materials like fibrin and collagen, along with innervation systems and stem cells. Reproduced with permission.^{[ 44 ]} Copyright 2017, Elsevier. C) Membrane formation, characterized by laminin and Thrombospondin‐1 expression in the corneal epithelial‐stromal interface layer. Reproduced under terms of the CC‐BY license.^{[ 43 ]} Copyright 2019, MDPI.

Recent studies have explored advanced co‐culture systems to foster direct cellular communication and basement membrane formation, as novel biomaterials and tissue‐engineered approaches alone have not fully replicated the native corneal microenvironment's complex cell‐to‐cell and layer‐to‐layer interactions. McKay et al. demonstrated the potential for replicating basement membrane formation (Figure 3C) between the corneal stroma and epithelial layer using co‐culture techniques with Transwell systems.^{[ 43 ]} This method is particularly noteworthy for directly layering epithelial cells over the self‐assembled corneal stroma, fostering cell‐to‐cell interaction without synthetic membranes. The enhanced layer‐to‐layer interaction resulted in laminin and Thrombospondin‐1 production, which is abundantly present in the basement membrane, providing critical evidence of membrane formations. Furthermore, Wang et al. developed an in vitro 3D corneal tissue model (Figure 3B) that included epithelial, stromal, and neuronal components.^{[ 44 ]} The model demonstrated the importance of cell–cell communication in maintaining corneal structure and function, as well as the potential for studying corneal innervation and neurotrophic keratopathy. By integrating tissue engineering approaches into cornea chips, there are many ways to improve current cornea chip models by incorporating multiple cell types and neural components, providing a more comprehensive representation of the cornea. However, they still face challenges in fully replicating the complexity of the native cornea. Issues such as achieving proper innervation, recreating the precise organization of collagen fibrils in the stroma, forming proper interfacial layers, and maintaining the long‐term viability of all cell types in co‐culture systems remain areas for future improvement.

1.3. Critical Challenges in Cornea Chip Development

Cornea chips face unique challenges compared to standard organ chips (such as lung, intestine, and heart valve chips). While other organ chips emphasize dynamic motions like cyclic stretching and strain, cornea chips must faithfully replicate specific features, including geometrical curvature, multilayered structure, optical transparency, and barrier integrity. Even though current cornea chip designs are evolving to incorporate the full spectrum of corneal properties more effectively, drawing upon advanced tissue engineering techniques, existing models simplify the complex architecture of the cornea due to constraints imposed by microfluidic channel designs.^{[ 36 ]} One of the challenges in corneal epithelium reconstruction is to recapitulate a multilayered and highly organized corneal epithelium consisting of basal cells, wing cells, and superficial squamous cells.^{[ 12} , ^{45 ]} Basal cells, which are cuboidal in shape and firmly attached to the underlying basement membrane, continuously divide and differentiate into wing cells that migrate toward the corneal surface.^{[ 14 ]} As these cells move anteriorly, they undergo further differentiation into flattened, terminally differentiated superficial squamous cells, which eventually lose their adhesion to the underlying layers and are shed into the tear film.^{[ 13 ]} These anatomical and physiological structures of the corneal epithelium should be explored to understand drug permeation and cornea dysfunctions.^{[ 46 ]} Another challenge lies in incorporating an orthogonally arranged stromal layer that mimics the native ECM composition and organization. The corneal stroma, which accounts for 90% of the corneal thickness, comprises highly organized collagen fibrils arranged in orthogonal lamellae, interspersed with proteoglycans and keratocytes.^{[ 47 ]} This unique ECM architecture is critical for maintaining corneal transparency, mechanical strength, and nutrient diffusion.^{[ 48 ]} Current cornea chip models often rely on hydrogels or scaffolds that fail to capture the complex ECM organization of the native stroma, leading to models with suboptimal optical and mechanical properties.^{[ 40 ]} Developing advanced biomaterials and fabrication techniques that enable the creation of anisotropic, multilayered stromal constructs with native‐like ECM composition will be a key step in advancing cornea chip technology.^{[ 44} , ^{49 ]} Incorporating corneal curvature into chip designs is a possible way to reconstruct native corneal stroma. The cornea curvature significantly regulates cell behavior and tissue morphogenesis, influencing cell migration, differentiation, and ECM remodeling.^{[ 50 ]} Neglecting this critical feature in cornea chip models may reduce physiological relevance and limit their predictive accuracy. Developing innovative fabrication techniques, such as 3D printing or molding, to create curved substrates that mimic the geometry of the cornea will be crucial for advancing the field.^{[ 42} , ^{51 ]} Moreover, integrating mechanical actuation systems, such as pneumatic or hydraulic pumps, into chip designs could enable the modulation of curvature to simulate various physiological and pathological conditions.^{[ 52 ]} Recently, Kim et al. demonstrated that a hydraulically controlled curvature chip significantly influenced quiescent corneal keratocytes, altering their phenotype, focal adhesions, and ECM remodeling, including changes in collagen and proteoglycans.^{[ 51b ]} Activated corneal fibroblasts also formed specific orthogonal patterns in response to curvature.^{[ 51b ]} This work presents the importance of incorporating curvature into corneal research to develop more physiologically relevant models.^{[ 51b ]} The novel in vitro platforms presented in this study offer valuable tools that can be further applied to corneal research. Another significant challenge in advancing cornea chip technology is integrating corneal endothelium. The endothelium, a monolayer of hexagonal cells, plays a crucial role in maintaining corneal hydration and transparency by regulating fluid transport and barrier function.^{[ 53 ]} Recreating this specialized cell layer in vitro requires careful optimization of culture conditions, including substrate stiffness, growth factor signaling, and fluid shear stress.^{[ 54 ]} Moreover, the endothelial cells form a functional barrier with tight junctions and express essential ion transport proteins, such as Na+/K+‐ATPase and aquaporins, to maintain corneal deturgescence.^{[ 20 ]} Developing strategies to promote the formation of a mature, functional endothelial layer will be essential for recapitulating biomimetic cornea chips.

Investigating the formation and function of the corneal basement membranes, namely BL and DM, is another important challenge in cornea chip development. These specialized ECM structures are critical in maintaining corneal integrity, providing structural support, and regulating cell adhesion and signaling.^{[ 55 ]} The BL, situated between the epithelium and stroma, is composed primarily of collagen type IV, laminin, and heparan sulfate proteoglycans.^{[ 55 ]} It serves as a substrate for epithelial cell adhesion and migration and plays a crucial role in wound healing and regeneration. DM, located between the stroma and endothelium, is a thick, acellular matrix composed of collagen type VIII and glycoproteins.^{[ 56 ]} It provides mechanical support to the endothelium and regulates fluid transport and barrier function. Recapitulating the structure and function of these basement membranes in cornea chip models requires a deep understanding of their composition, organization, and biogenesis. Co‐culture systems that enable the direct interaction between epithelial, stromal, and endothelial cells will be essential for promoting the formation of physiologically relevant basement membranes.^{[ 57 ]} Moreover, the use of biomimetic materials, such as silk fibroin or decellularized corneal ECM, as substrates for cell growth and differentiation may facilitate the deposition and assembly of native‐like basement membrane components.^{[ 58 ]}

Incorporating corneal innervation into the stroma and epithelium is another crucial challenge in developing native‐like cornea chips. The cornea is one of the most densely innervated tissues in the body, with sensory nerves originating from the trigeminal ganglion.^{[ 59 ]} These nerves form a complex network within the stroma and epithelium, with nerve endings extending into the superficial epithelial layers. Corneal innervation plays a vital role in maintaining corneal sensitivity, promoting epithelial wound healing, and regulating tear production.^{[ 60 ]} After corneal transplantation, the recovery of the neuronal system is a gradual process that can take several months to years. The severed corneal nerves undergo a process of regeneration, which involves the growth of new nerve fibers from the remaining nerve bundles at the periphery of the cornea towards the center.^{[ 61 ]} However, this process is often incomplete, and the density and organization of the regenerated nerves may not fully restore the preoperative innervation levels.^{[ 62 ]} To address this challenge in cornea chip models, it is possible to implement innervation by adding a perimeter or separate chamber that allows for the growth and integration of functional nerve fibers into the stroma and epithelium, similar to these chip designs.^{[ 63 ]} The perimeter chamber approach involves creating a separate compartment surrounding the main cornea chip structure, which can be seeded with sensory neurons or dorsal root ganglion (DRG) explants. This compartment can be connected to the main cornea chip through microchannels or porous membranes, allowing for the guided growth and extension of nerve fibers into the stromal and epithelial layers. Advanced biomaterials and fabrication techniques can be employed to facilitate the integration of nerve fibers into the cornea chip. For example, the use of aligned nanofiber scaffolds or micropatterned substrates can provide topographical cues for guided neurite outgrowth and help recreate the complex neural architecture of the native cornea.^{[ 64 ]} Moreover, the incorporation of hydrogels or ECM‐derived materials within the perimeter chamber can provide a supportive microenvironment for nerve fiber growth and promote the formation of functional synaptic connections with corneal cells.^{[ 65 ]} By incorporating a perimeter chamber for corneal innervation, researchers can develop cornea chip models that more closely recapitulate the native neuronal system and its recovery process after corneal transplantation. This approach will enable the investigation of nerve–cell interactions, the role of innervation in corneal homeostasis, and the evaluation of potential therapeutic strategies for promoting nerve regeneration and restoring corneal sensitivity in various physiological and pathological conditions.

1.4. Cornea Disease Models

Corneal blindness, a significant global health challenge, primarily stems from epithelial injuries caused by bacterial, fungal, and microbial infections, along with burns and scarring from external factors.^{[ 66 ]} Despite the growing need for effective treatments, developing therapies for corneal diseases still needs to be improved due to the lack of suitable in vitro models and the limitations of animal and ex vivo studies, which often fail to replicate the human corneal environment. Current in vitro disease modeling using cornea‐on‐chips faces significant challenges due to the complexity of wound healing mechanisms in the corneal stroma following epithelial injury.^{[ 67 ]} The absence of fully integrated cornea‐on‐chip models that incorporate critical elements, such as innervation and immune systems, also hinders the advancement of pharmacological and surgical treatments for corneal repair. This section discusses the cornea disease models for four major corneal diseases: infectious keratitis, dry eye disease, keratoconus, and Fuchs' dystrophy.

Infectious keratitis, particularly bacterial keratitis, has seen a rising incidence due to the widespread use of contact lenses. The primary culprits in these infections are well‐known pathogens such as Staphylococcus aureus (SA), Pseudomonas aeruginosa (PA), and Streptococcus pneumoniae (SP).^{[ 68 ]} These bacterial invaders can lead to severe corneal complications, including perforation, extensive tissue damage, and pronounced alterations in the shape of the corneal stroma. The resulting symptoms often manifest as corneal swelling, opacity, and ultimately compromised vision. Traditional treatment with antibiotic eye drops is becoming less effective due to the growing problem of antibiotic resistance.^{[ 69 ]} This challenge has spurred research into novel pharmacological treatments, including drug‐eluting contact lenses and corneal bandages. Recent advancements have led to the development of a corneal keratitis model induced by SA on a microfluidic chip to test the efficacy of three antibiotics: Chloramphenicol (CPL), Tobramycin (TOB), and Levofloxacin (LFX). Deng et al. successfully demonstrated drug efficiency, with bacterial counts reduced to 40 CFU mL⁻¹ with TOB and 72.5 CFU mL⁻¹ with LFX, compared to the control (2.51 × 10⁹ CFU mL⁻¹) and CPL (1.06 × 10⁵ CFU mL⁻¹).^{[ 70 ]} This chip could be improved by incorporating BL, which would provide an additional barrier against bacterial penetration and potentially enhance cell viability. Additionally, the model lacks sufficient investigation into ECM remodeling and fibrosis activation within the stromal layer, which are critical for fully recapitulating the pathological progression of corneal keratitis. In comparison, tissue‐engineered approaches have demonstrated more comprehensive analyses of keratitis progression. Marina et al. utilized a reconstructed 3D corneal stroma model, embedding human corneal fibroblasts within a collagen hydrogel in a Transwell system.^{[ 71 ]} This study revealed that fungal penetration upregulated collagenase expression, triggering a pro‐inflammatory response and inducing morphological changes in corneal tissue. Such detailed analysis, including ECM remodeling and inflammatory responses, highlights the need for similar comprehensive characterization in corneal keratitis chip studies to enhance their physiological relevance and accuracy. In addition, Brackman et al. developed Hamamelitannin (HAM) and various quorum sensing inhibitors to observe the bacterial response to antibiotic susceptibility in SA biofilms.^{[ 72 ]} This work demonstrated that HAM enhances the susceptibility of SA biofilms to antibiotics by disrupting peptidoglycan biosynthesis and extracellular DNA release. This work provides insight into how corneal keratitis models can be used to test novel treatments, potentially improving therapeutic outcomes and drug efficiency. However, these in vitro keratitis models have limitations in accurately determining suitable infection times and inoculation doses.^{[ 73 ]} By integrating tissue engineering models with organ chip platforms, these limitations can be addressed by utilizing the advantages of organ chips, which can precisely control flow rates, exposure times, and drug concentrations. The chip can simulate physiological conditions more accurately and facilitate the establishment of optimal infection and treatment parameters. Moreover, integrating high‐throughput screening systems into corneal keratitis chips allows large‐scale testing of diverse drug concentrations and novel treatments while maintaining precise control over infection conditions, such as bacterial load and exposure timing. However, such systems require advanced automation to ensure reproducibility and efficiency when managing high‐throughput experiments. To enhance further, current organ chips have incorporated non‐invasive TEER measuring systems by creating TEER measurement ports or embedding electrodes within each microfluidic layer.^{[ 74 ]} While TEER provides real‐time analysis of barrier function, high‐throughput keratitis models will benefit from upgraded systems equipped with multiple electrodes to enable simultaneous, real‐time measurements across different regions and time points.^{[ 75 ]}

DED affects 5–35% of the population and is characterized by insufficient tear production or rapid tear evaporation, primarily impacting the corneal tear film and epithelium.^{[ 76 ]} Several studies have developed DED models using microfluidic cornea chips that successfully reconstructed multilayered cornea epithelium, including tear film, by applying air‐lifting motion.^{[ 9} , ³¹ , ^{32 ]} However, most existing models focus solely on the epithelium and tear film. Kheirkhah et al. demonstrated that severe DED is associated with a significant decrease in sub‐basal corneal nerve density and central corneal endothelial cell density.^{[ 77 ]} This suggests proper DED models should include corneal stroma and endothelium layers to study their effects on nerve function and corneal hydration. Understanding the relationship between DED, endothelium, and innervation could significantly enhance the evaluation of novel treatments, offering a more reliable platform for testing drug efficacy and optimizing therapeutic strategies. Additionally, Weber et al. suggest that scleral contact lenses, which create a saline‐filled space between the lens and corneal surface, could improve severe DED treatment.^{[ 78 ]} Future in vitro DED models could expedite the development of new treatments, including drug‐infused lenses and therapies like scleral contact lenses, while minimizing reliance on animal and human models.

Keratoconus is a progressive corneal disorder that primarily weakens and thins the stromal layer of the cornea. This thinning causes the cornea to bulge outward into a cone‐like shape, leading to three main visual complications: distorted vision, increased sensitivity to light, and reduced visual clarity.^{[ 79 ]} The condition can be associated with post‐vision correction surgeries and hereditary factors. Abnormal collagen and proteoglycan expression in the corneal stroma, driven by matrix metalloproteinase activation, leads to the curvature deformation characteristic of keratoconus.^{[ 80 ]} In vitro, keratoconus models have been developed by transforming quiescent corneal keratocytes into activated form and using induced pluripotent stem cells (iPSCs).^{[ 81 ]} These models successfully formed self‐assembled corneal stromal structures over a 4‐week culture, simulating the keratocyte transformation seen in keratoconus. However, they still lack the ability to fully replicate the 3D structure with the cell density gradient of the native corneal stroma and still have limitations in incorporating it into the organ chip format. Many studies overlook the critical role of curvature in modulating ECM structures within the corneal stroma. Pieuchot et al. (2018) demonstrated that curvature‐induced traction forces influence cell migration and behaviors related to focal adhesions.^{[ 82 ]} Based on the review of multiple studies,^{[ 83 ]} curvature, as demonstrated by its influence on cell alignment, may also play a role in activating matrix metalloproteinases, which are critical factors in developing keratoconus. Kim et al. developed a hydraulically controlled curvature chip to investigate corneal stromal cell behaviors under various curvatures that mimic corneal ectasia diseases.^{[ 51b ]} In keratoconus and keratoglobus cases with curvatures exceeding 45 diopters, quiescent corneal keratocytes show significant phenotypic changes, including decreased ALDH3 expression, increased α‐SMA expression, and altered ECM arrangement and composition—all changes typically observed during wound healing processes. Leveraging these insights, future keratoconus models should prioritize the in‐depth study of curvature effects, as simulating diverse curvatures on microfluidic chips could reveal critical mechanisms for new treatments. For instance, Sharif et al. demonstrated that collagen cross‐linking treatment, typically tested on animals or humans, can be effectively evaluated using in vitro keratoconus models.^{[ 84 ]} Building on the promising results of the Bioengineered Porcine Construct Double‐crosslinked, which has shown significant potential in both clinical and in vivo models, a similar approach could be applied to in vitro keratoconus models.^{[ 85 ]} This approach could incorporate novel treatments, such as dexamethasone for sustained anti‐inflammatory activity, to further assess efficacy and therapeutic potential without relying on animal models.^{[ 86 ]} Looking ahead, an integrated keratoconus model that incorporates curvature effects could serve purposes: evaluating the effectiveness of emerging treatments and uncovering the factors that trigger keratoconus through various mechanosensitive signaling pathways. Such a model would provide a comprehensive platform for advancing our understanding and treatment of this complex corneal disorder.

Fuchs endothelial corneal dystrophy (FECD), the most common primary corneal endothelial dystrophy, represents the leading indication for corneal transplantation worldwide.^{[ 87 ]} This genetic disease is characterized by progressive corneal endothelial cell decline, resulting in apoptosis, cellular morphological changes, and abnormal ECM formation. The endothelial cell loss of FECD results in the loss of barrier function and the inability of the cells to maintain fluid balance due to the functional loss of the Na+/K+ pump.^{[ 88 ]} Although the mechanisms of FECD are not fully understood, they involve osmoregulation loss, edema, and cell death, leading to corneal swelling and vision impairment—pathological swelling of the corneal endothelium results in loss of transparency and vision problems. While significant efforts have been made to develop new therapeutic methods and drugs for FECD, evaluating and screening their efficacy remains challenging.^{[ 89 ]} Developing a proper FECD model remains crucial for advancing treatment and understanding FECD mechanisms. One major challenge of an engineering model of FECD is to maintain corneal endothelial cell phenotypes and functions in vitro, where cells often exhibit irregular shapes and compromised tight junctions, unlike their in vivo endothelial cells. Palchesko et al. highlighted the sensitivity of corneal endothelial cells to substrate stiffness, showing that cells cultured on substrates mimicking the physiological stiffness of DM (50 kPa) had superior morphology.^{[ 90 ]} This underscores the importance of mechanical properties in endothelial cell culture. However, the effects of curvature, which provides mechanical cues by varying substrate stiffness, have yet to be fully explored. In their initial study, Kim et al. observed that corneal endothelial cells respond to changes in curvature by altering nuclear size, proliferation rate, and shape while maintaining ion channel function.^{[ 91 ]} Ongoing studies are now investigating the deeper impact of curvature on ion channel activity using advanced techniques such as ion pump sensors and molecular‐level analysis. The outcomes of this research are expected to enhance our understanding of the underlying mechanisms and contribute to the development of more accurate in vitro models that better replicate corneal endothelial function. Complementing these curvature studies, researchers are exploring additional approaches to improve corneal endothelial cell models. For instance, Fuchs dystrophy can be modeled in vitro by exposing the endothelial cells to Ultraviolet A (UVA) to induce DNA damage and generate oxidative stress.^{[ 92 ]} In a related study, Liu et al. developed a nongenetic animal model to replicate the FECD by exposing mouse corneas to UVA light. The study found that UVA exposure led to increased reactive oxygen species production, causing significant corneal endothelial cell loss, particularly in female mice.^{[ 93 ]} Additionally, the use of progenitor cells derived from iPSCs in cornea chips shows promise in creating healthier, hexagonal endothelial cells in vitro.^{[ 94 ]} However, the isolation and culture of patient‐derived endothelial cells—both diseased and healthy—presents notable challenges, including cell separation, low cell proliferative capacity of adult CECs, and the difficulty of maintaining phenotypic stability during prolonged in vitro culture. Overcoming these limitations requires optimized isolation protocols,^{[ 95 ]} improved culture systems,^{[ 96 ]} and advanced differentiation strategies such as the use of curvature substrate and iPSCs. Despite these challenges, the integration of patient‐driven cells with the cornea chip technology can provide valuable tools for developing customizable drugs and treatments. This combination of advanced modeling techniques and patient‐specific cells represents a significant step towards personalized medicine in corneal research and treatment.

1.5. Conclusion and Future Directions

Cornea chip technology marks a significant advancement in ocular research, providing a powerful platform for investigating corneal biology, disease mechanisms, and drug development. These chips replicate the complex structure and function of the human cornea in vitro, bridging the gap between traditional animal models and human clinical studies. This provides researchers with a physiologically relevant and predictive tool for both basic and translational research. As highlighted in this review, significant progress has been made in the development of cornea chip models that mimic various aspects of the native tissue, including the multilayered epithelium, stroma, and endothelium. These models have been used to study a wide range of corneal phenomena, from drug permeability and toxicity to wound healing and disease pathogenesis. However, several key challenges remain in creating truly biomimetic cornea chips, including the incorporation of native ECM composition and organization, the replication of corneal curvature, and the formation of functional basement membranes.

Several promising developments in cornea chip technology are poised to transform ocular research (Figure 4 ). As discussed earlier, cornea curvature plays a critical role in regulating cell behavior and tissue morphogenesis, and its incorporation into chip designs is essential for creating genuinely biomimetic models. Our research group is developing a novel hydraulically controlled cornea chip to mimic the cornea curvature and investigate its effects on cell behavior and interfacial membrane formation.^{[ 51b ]} The hydraulically controlled cornea chip consists of a multilayered PDMS structure with embedded microfluidic channels that enable the precise control of curvature through fluid injection. By varying the fluid pressure, we can simulate various physiological and pathological curvatures, such as those observed in corneal ectasia. The chip can also incorporate a biodegradable Poly (lactic‐co‐glycolic acid) membrane that serves as a substrate for epithelial cell growth and differentiation while allowing direct interaction with the underlying stromal layer.^{[ 97 ]} This design enables us to investigate the effects of curvature on cell migration, differentiation, and ECM remodeling, as well as the formation of the epithelial basement membrane.

Figure 4.

Future directions for cornea chip technology, highlighting advancements in structure, function, design, materials, integration with other ocular tissues, and personalized medicine approaches.

Another key future direction is the development of cornea chips that incorporate immune cells and blood vessels to study corneal inflammation and angiogenesis. The cornea is an immune‐privileged tissue with a complex network of resident and infiltrating immune cells that play crucial roles in maintaining homeostasis and responding to injury or infection. Dysregulation of the corneal immune response can lead to chronic inflammation, neovascularization, and loss of transparency, as observed in conditions such as keratitis, corneal graft rejection, and limbal stem cell deficiency. Developing cornea chips that incorporate immune cells, such as macrophages, dendritic cells, and T cells, could provide valuable insights into the mechanisms underlying corneal inflammation and enable the testing of novel immunomodulatory therapies. Moreover, the integration of microfluidic channels that mimic the limbal blood vessels could allow for the study of angiogenic factors and the screening of anti‐angiogenic drugs. Such models could also be used to investigate the interplay between inflammation and angiogenesis in the context of corneal wound healing and regeneration.

Patient‐derived cells and iPSCs represent another promising direction in cornea chip development. Creating personalized cornea chips with patient‐specific cells could revolutionize precision medicine, enabling tailored therapies for corneal diseases. Patient‐derived primary cells, such as limbal stem cells, keratocytes, and endothelial cells, are directly isolated from corneal tissues and can be incorporated into the cornea chip. These cells allow researchers could study the patient‐specific genetic and epigenetic factors contributing to disease susceptibility and progression. However, primary corneal cells face limitations, including the difficulty in maintaining cell‐specific phenotype, low proliferative capacity, and limited passaging availability during long‐term in vitro culture. In contrast, iPSCs can be efficiently differentiated into progenitor corneal epithelial, stromal, and endothelial cells using defined protocols that recapitulate developmental pathways. iPSC‐derived cells address critical limitations of primary human corneal cells by providing progenitor properties and regenerative potential, including enhanced proliferation capacity, stable phenotype maintenance, and the ability to self‐renew. Both patient‐derived and iPSC‐derived cornea chips could serve as powerful tools for drug screening and toxicity testing, allowing for the identification of the most effective and safe therapies. iPSC‐derived cells are particularly advantageous for long‐term studies and therapeutic applications because they reprogram the senescence^{[ 98 ]} and overcome limited expansion capacity^{[ 99 ]} seen with primary cells, while maintaining normal karyotype and functional properties across multiple passages. iPSC‐derived cornea chips offer three key advantages for transplantation research and regenerative medicine: they provide a source of immunologically matched tissue, allow genetic modification for enhanced therapeutic properties, and enable the study of patient‐specific disease mechanisms.^{[ 100 ]} This is especially valuable for investigating inherited corneal dystrophies and developing targeted treatments for conditions where conventional therapies have limited efficacy.^{[ 101 ]}

The development of multiorgan chip systems that integrate cornea chips with other ocular tissue models, such as the retina, choroid, and lens, is another exciting future direction. The eye is a complex organ with intricate interactions between its various components, and many ocular diseases involve multiple tissues. For example, diabetic retinopathy, a leading cause of blindness, affects both the retina and the cornea, leading to corneal nerve degeneration and delayed wound healing. Similarly, uveitis, an inflammatory condition of the uveal tract, can cause secondary corneal complications, such as band keratopathy and corneal edema. Multiorgan chip systems that model interactions between different ocular tissues can enhance our understanding of eye physiology and pathology at a systemic level. Such models could enable the study of the molecular and cellular mechanisms underlying ocular diseases and facilitate the development of targeted therapies that address the multitissue nature of these conditions. Moreover, multi‐organ chip systems could serve as powerful tools for drug delivery and pharmacokinetic studies, allowing for the optimization of drug formulations and dosing regimens.

Lastly, the integration of advanced sensing and imaging technologies into cornea chip designs is a key future direction that could significantly enhance their utility and translational potential. The development of non‐invasive, real‐time monitoring systems that can assess corneal structure, function, and cellular activity could provide valuable insights into the dynamic processes occurring within the chip. For example, Kim et al. developed an electroretinogram (ERG) capable of detecting and recording bioelectrical signals from retinal cells in response to light.^{[ 102 ]} This work demonstrates the potential for integrating advanced biosensors into cornea‐on‐a‐chip systems. Similarly, the incorporation of biosensors and reporter systems that can detect changes in specific analytes, such as glucose, lactate, or inflammatory cytokines, could provide real‐time information on the metabolic and pathophysiological states of the cornea chip. Such technologies could enable the early detection of adverse events, such as drug toxicity or infection, and allow for the timely intervention and optimization of treatment strategies. Moreover, the integration of advanced sensing technologies could facilitate the development of high‐throughput screening platforms that can rapidly assess the efficacy and safety of novel therapies, accelerating the drug discovery and development process.

Conflict of Interest

The authors declare no conflict of interest.

Biographies

Minju Kim is a Ph.D. candidate in the Department of Mechanical Engineering at the University of Utah. She received her B.S. and M.S. degrees in Biomedical Engineering. During MS, she studied mesenchymal stem cell differentiation and endothelial cell interactions using in vitro biomimetic cell culture systems. Her current research focuses on how curvature acts as a biomechanical cue regulating corneal cell behavior, integrating organ‐on‐a‐chip technology. She specializes in microfabrication, biomaterials, and mechanobiology to develop biomimetic platforms for modeling corneal ectasia. By unraveling curvature‐driven cellular responses, she aims to advance in vitro disease modeling for ophthalmic research.

Kanghoon (Kyle) Choi is a Ph.D. candidate in Mechanical Engineering at the University of Utah. He received B.S. degrees in Biomedical Engineering and M.S. degrees in Mechanical Engineering at the University of Utah. His research focuses on developing novel hydrogel‐integrated microfluidic platforms to study how trabecular meshwork cells respond to combined mechanical forces including substrate stiffness, shear stress, and tensile stress. He specializes in microfabrication, finite element analysis, and mechanobiology to develop biomimetic platforms for studying trabecular meshwork cell behavior. Through investigating the complex interplay of mechanical stresses in trabecular meshwork cells, he contributes to accelerate the development of effective glaucoma therapeutic strategies.

Amy Lin is Professor of Ophthalmology at the University of Utah Moran Eye Center. She is a cornea specialist and is medical director of Utah Lions Eye Bank.

Jungkyu (Jay) Kim is an Associate Professor in the Department of Mechanical Engineering at the University of Utah, with adjunct appointments in Biomedical Engineering and the Moran Eye Center. After earning his Ph.D. in Biomedical Engineering from Utah and completing postdoctoral training at UC Berkeley, he served as an Assistant Professor at Texas Tech University. His research focuses on point‐of‐care diagnostics using microfluidic platforms, organ‐on‐chip models of ocular and renal systems, and microfabricated instruments for space exploration. His work is supported by NSF, DOE, NIH, and NASA.

Kim M., Choi K., Lin A., Kim J., Current and Future Cornea Chip Models for Advancing Ophthalmic Research and Therapeutics. Adv. Biology 2025, 9, 2400571. 10.1002/adbi.202400571