Modeling early pathophysiological phenotypes of diabetic retinopathy in a human inner blood-retinal barrier-on-a-chip

Abstract

Diabetic retinopathy (DR) stands as a pervasive and debilitating microvascular complication of diabetes, globally recognized as a leading cause of irreversible vision loss and blindness among the working-age population. At its pathological core, DR is characterized by the progressive deterioration of the inner blood-retinal barrier (iBRB), a highly specialized and exquisitely selective neurovascular unit essential for maintaining retinal homeostasis and normal visual function. The integrity of the iBRB is crucial for regulating the passage of molecules, nutrients, and immune cells between the blood and the delicate retinal tissue, thereby protecting the neural retina from circulating toxins and inflammatory mediators. When this barrier is compromised, it leads to vascular leakage, edema, and ultimately, severe damage to the retinal structure, culminating in profound and often irreversible visual impairment. While the overt clinical symptoms of DR are well-documented, the precise underlying disease mechanisms—including hallmark features such as basement membrane thickening, the selective dropout of pericytes, and widespread capillary damage—remain incompletely understood. This lack of comprehensive mechanistic insight has significantly hampered the development of effective, targeted interventions specifically designed to repair and restore the diseased iBRB microvascular networks, leaving a considerable unmet clinical need for novel therapeutic strategies that go beyond current, often palliative, treatments.

A significant hurdle in advancing our understanding of DR pathogenesis and discovering new therapeutic targets has been the inherent limitations of conventional research models. Traditional animal models, while providing invaluable *in vivo* insights, often fail to fully recapitulate the complex nuances of human DR due to species-specific physiological differences, ethical considerations, and the inherent challenges in real-time, high-resolution observation of microvascular changes. Similarly, simplistic *in vitro* systems, while offering controlled environments, frequently lack the crucial physiological complexity, dynamic flow conditions, precise cellular interactions, and the three-dimensional architectural fidelity necessary to truly mimic the intricate retinal microvasculature. Consequently, these models often suffer from a critical lack of translatability and predictive power, making it challenging to reliably identify novel target pathways or screen potential drug candidates that will prove effective in a clinical setting.

Addressing these critical limitations, we present here the innovative development of a novel biomimetic platform: a diabetic inner blood-retinal barrier-on-a-chip. This sophisticated microfluidic device has been engineered with meticulous attention to detail, precisely mimicking the complex architecture and physiological conditions of the human retinal microvasculature. Its design enables the recreation of key pathophysiological phenotypes and disease pathways *in vitro* that are remarkably representative of those observed in clinical diagnoses of diabetic retinopathy. This advanced model bridges the gap between simplified *in vitro* systems and complex *in vivo* models, offering an unparalleled tool for mechanistic studies and drug discovery.

Through systematic experimentation, we conclusively demonstrate that diabetic stimulation of our iBRB-on-a-chip successfully mirrors several core pathological features characteristic of human diabetic retinopathy. The model faithfully recapitulates the devastating phenomenon of pericyte loss, a critical event in DR where these mural cells, essential for vascular stability and integrity, detach from the retinal capillaries, leading to an unstable and leaky vasculature. This pericyte dropout is followed by observable vascular regression, a process involving the actual collapse and loss of retinal capillaries, resulting in areas of non-perfusion and retinal ischemia. Furthermore, the chip model displays the formation of “ghost vessels,” which are acellular capillary remnants, providing a clear histological hallmark of previous, now compromised, vasculature. Beyond structural changes, the diabetic iBRB-on-a-chip also accurately reflects the chronic inflammatory environment typical of DR by demonstrating the sustained production and release of various pro-inflammatory factors, which are known to contribute to iBRB breakdown and neurodegeneration in the retina. The faithful replication of these diverse and clinically relevant features provides strong validation for the utility and translatability of our engineered platform.

In addition to replicating key morphological and inflammatory hallmarks, our study generated comprehensive transcriptomic data from the diabetic iBRB microvascular networks established on the chip. This high-resolution gene expression analysis provides an invaluable molecular blueprint of the cellular responses to diabetic stress within the iBRB. By identifying specific genes and pathways that are aberrantly expressed or dysregulated in this diabetic microenvironment, Lanraplenib these transcriptomic datasets hold immense potential to reveal previously unappreciated drug targets, offering novel molecular avenues for therapeutic intervention. Furthermore, leveraging the unique capabilities of our chip model, we rigorously examined various pericyte-endothelial cell stabilizing strategies. This investigation focused on approaches aimed at preserving the crucial symbiotic relationship between pericytes and endothelial cells, a partnership that forms the structural and functional basis of a healthy iBRB. By exploring different interventions, we aim to identify compounds or biological agents that can strengthen or repair the compromised barrier, thereby preventing or reversing the pathological leakage and subsequent vision loss in DR.

In summary, the diabetic inner blood-retinal barrier-on-a-chip developed in this study represents a significant technological advancement in the field of DR research. Our model convincingly recapitulates key pathological features and intricate molecular mechanisms of the disease, providing a highly relevant and controllable *in vitro* system. This innovative platform holds substantial promise to inform and accelerate the discovery and development of future therapeutic strategies for diabetic retinopathy. By offering a more accurate and predictive model for investigating disease progression and testing novel interventions, it paves the way for a new generation of targeted therapies designed to preserve retinal microvascular integrity and ultimately prevent irreversible vision loss in patients suffering from this widespread and devastating complication of diabetes.

Keywords: Diabetic retinopathy; Inner blood-retinal barrier; Organ-on-a-chip; Pericytes; Vascular regression; Transcriptomics; Drug discovery.