In Silico Is Not a Lonely Option: In Vitro Organoids Test New Pharmaceuticals

In 2014, I remarked here about in silico modeling for digital biology. The relationship between AI-driven digital and in vitro modeling in systems biology is developing to some extent in synch and in a mutually cross-fertilizing manner for the purpose of minimizing both clinical and animal testing while expediting, improving and reducing cost of biomedical innovation. Here is a sketch of how this might work in a more organized fashion:

Organoids are artificial organ-like structures a few millimeters in size bred in vitro that show realistic micro-anatomy. They are developed to perform medical tests for which otherwise animal testing would have to be used, that, ethical issues aside, is neither a very close nor highly reliable approximation of human organs. Such partial models of the brain, heart or kidneys have already been created in recent years for narrowly-tailored specific purposes. Compared to in silico models, they involve genuine human tissue and higher levels of complexity in some respects.

Researchers working with Josef Penninger at the Institute of Molecular Biotechnology (IMBA) in Vienna, Austria now report that they took a further step: they established 3D tissue cultures that permit testing blood vessels on the micro scale. Among additional future uses, their model presented in Nature serves to simulate diabetic vascular damage.

Diabetics frequently face vascular diseases as a consequence of their systemic condition. In the case of major blood vessels, potential consequences include faster occurrences of atherosclerosis resulting in myocardial infarction, stroke, and amputation of the lower limbs. However, diabetic damage affecting small vessels, micro-vasculopathy, can have similarly catastrophic results as it triggers diabetic kidney disease that requires many diabetics to undergo regular dialysis, and/or causes diabetic retinopathy. Work on a 3D vascular organoid model was finalized by a team of Reiner Wimmer, who created 1-2-millimeter sized models of human capillaries, complete with endothelial cells as a cavity lining layer as well as pericytes as supporting cells.

To create a vascular model of capillaries, the researchers induced human stem cells to differentiate into functional blood vessels of a micro-format. These human blood vessel organoids were then implanted into the kidneys of immunodeficient mice that do not reject human cell tissue. The organoids grew to form a stable, perfused vascular tree including arteries, arterioles and venules. In 95 percent of cases, the organoids survived longer than six months. This experimental design also appears to be well suited for research on microvascular disease resulting from diabetes mellitus type 2. In this type of microvascular disease, one essential feature is the thickening and onion-like structure of these vessels’ basement membrane and loss of vascular cells. The endothelial cell layer “leaks,” and such leakage leads in diabetics to retinal disease that causes a typical damage to the ocular fundus ending in blindness. Diabetes impairs the functions of endothelial cells and disturbs communication between the endothelium and pericytes.

The experiment imitated type 2 diabetes in the lab by exposing the blood vessel organoids to hyperglycemia by inducing high concentrations of sugar and inflammatory messengers (cytokines), which in vitro induces thickening of the vascular basement membrane. Down to the level of electron microscopy, all the typical pathological changes to the capillaries became visible.  The authors derived a first set of results from tests in which the diabetic organoids were treated with a wide range of conventional antidiabetic medications aimed at lowering blood sugar levels. None of them showed any effect on the changes occurring in the vascular organoids. Perhaps the best result came from an inhibitor of gamma-secretase enzyme (DAPT). DAPT reduced the formation of vascular thickening type IV collagen and normalized the emergence of endothelial cells. According to the results of first impression, this is apparently due to an effect attributed to the expression of the NOTCH3 gene. Gamma-secretase inhibitors have so far been tested as potential Alzheimer drugs. This may yet prove important to future search for new strategies for prevention and treatment of diabetic vascular disease.

The exciting part about this research is that the authors were able to produce genuine human blood vessels out of stem cells. These organoids show very high similarity to human capillaries, which permits conducting meaningful studies of vascular diseases directly in live human tissue. This study in experimental biology proved that organoids derived from human stem cells faithfully reproduce the structure and function of human blood vessels and are useful systems for modelling and identifying the regulators of diabetic vasculopathy, a disease that affects hundreds of millions of patients worldwide. To the extent the data from such models can be fed into AI to improve in silico modeling, a significant acceleration of pharmacological testing cycles seems to be within reach.