The future of cerebral organoids in drug discovery

Recapitulating human development has been the central doctrine of the developmental biology field. Since the initial discovery of both human embryonic stem cells (ESCs)and human induced pluripotent stem cells (hiPSCs)there has been a multitude of studies that define the chemical signals to push forward human pluripotent stem cells into different states.

Many early applied differentiation experiments mimicked early developmental biology by transitioning through a spheroid intermediate known as an embryoid body (EB). EBs have spontaneous differentiation toward all three germ layers.

The 3 germlayers are isolated and subsequent specific enrichment of cell types such as neurons, cardiomyocytes, or hepatocytes can be achieved by using cell isolation methods and specific media that contain protein or small molecules that push the cells down specific developmental pathways. Initial attempts at producing aggregated models of brain development were largely conducted using mouse embryonic stem cells in the mid-2000′s.

Adaptation of these models to human systems then arose a few years later. Soon thereafter, it was found that added patterning factors are not necessary; rather, protocols that utilize a natural propensity of PSCs to initiate corticogenesis were developed.

In these models, PSCs are cultured as aggregates in a minimalistic medium, which leads to the formation of spontaneous 3D neuroectoderm structures within these aggregates. These spontaneous 3D structures adopt the architecture of a miniature immature human brain with many of the cell-types and the overall organization.

Adoption in biopharma for drug discovery

The majority of cerebral organoid studies have been conducted for basic research in academia and have not been adopted by the biopharma industry for drug discovery. In part, industry is more conservative and wants to see proof of value before making substantial investments in a nascent technology.

Newer technologies being implemented in industry must face the same raised level of scrutiny as established ones, as these campaigns ultimately lead to treatments for patients in addition to general scientific knowledge. There are a few notable examples in industry, however, where organoids have been used for drug and target discovery.

While there may be additional efforts in biopharma companies that are not highlighted here, this section will cover a few of the efforts that have been publicly disclosed. Stemonix is a service provider and provides assay ready microbrain organoids derived from iPSCs as a product for drug screening.

The company claims to have industrialized the process. However it is not clear the exact method they use or how standardized and reproducible their system is. The main focus is to provide organoids for toxicity testing that could be used to supplement the standard drug discovery process. However, there are a few examples of this platform being used to model diseases such as Rett Syndrome and to simulate viral infection.

The team also provides a custom discovery service and will work with clients to set up drug discovery projects. While it is clear that Stemonix has created a business out of providing organoids as a service and is generating income based on them it is not clear how many other companies use their services and buy their products.

System 1 biosciences is a neurotherapeutics company that employs AI-driven phenotypic screening to discover novel drugs for complex neurological and psychiatric diseases such as epilepsy, autism, and schizophrenia, disorders for which current discovery techniques have proved least successful. System 1 is employing iPSC-derived cerebral organoids for phenotypic-based drug discovery for psychiatric diseases where there is a large unmet medical need.

It is not clear if there are positive outcomes and whether drugs will come out of their pipeline in the near term. On average it takes more than 10 years for drugs to go from discovery to approval, and it is the hope of several of these companies to accelerate this process with improved models. We are going to have to wait to see if their approach will bear any fruit. a:head bio is an extremely new startup that has formed via a partnership with the Institute of Molecular Biotechnology (IMBA) in Vienna, Austria.

While new to the scene, a:head has the advantage of several years of cerebral organoid development through its close ties with Knoblish and Lancaster of IMBA. a:head’s focus is on the development of new therapeutics in the CNS space, with perhaps an initial focus on Dravet syndrome, a severe epilepsy caused by loss-of-function mutations in the sodium channel gene SCN1A.

Potential technical solutions to the challenges with cerebral organoids

Systems such as the Hamilton Star and Vantage are liquid handling robots that have been adopted for automated cell culture. These systems can be integrated with CO2 incubators and imaging platforms to generate fully integrated end-to-end cell culture systems that can maintain, passage and differentiate stem cells.

These systems could be adapted to generate uniform cerebral organoids in an automated manner to take out the variability introduced by multiple human hands. This could greatly standardize the generation of uniform cerebral organoids and take out the manual human variables that plague current organoid generation.

The neuroscience team at NIBR has two of these platforms and has been adapting the systems to grow cerebral organoids at scale. While it is still early and it has yet to be determined if these systems can improve organoid generation, it seems likely that other groups in industry will adopt these methods and continue to develop methods that improve the protocols and methods to reproducibly generate cerebral organoids.

CRISPR has revolutionized functional genomics, making it possible to generate isogenic iPSC lines for disease modeling with unprecedented speed, flexibility, and precision. This has the potential to greatly increase the reproducibility of disease models and has the opportunity to allow the identification of subtle phenotypes in cerebral organoids.

While the generation of isogenic disease models has become increasingly prevalent in two-dimensional cultures and other tissue types, there are only a few examples of gene editing being used for disease modeling in cerebral organoids. While none of these models have been used for large scale high-throughput screens yet, they have been used to identify the receptors for Zika virus infection.

The team at NIBR generated a high-throughput system to generate gene knockout in human pluripotent stem cells and to rapidly turn these into neuronal precursor cells and neurons. This system relies on a doxycycline inducible CAS9 and the introduction of single gRNAs via lentivirus. This system was used to efficiently knock out the punitive Zika receptor AXL and demonstrate that it was not required for infection.

Knock out of AXL also didn’t appear to inhibit cerebral organoid generation. This kind of isogenic modeling helps researchers to work around the differentiation variability that occurs when different cell lines are differentiated into cerebral organoids. Researchers can be more confident in these systems that the phenotype they are observing is indeed caused by the specific perturbation that they are introducing.

Microscopic dissection of organoids holds great promise for understanding developmental neurobiology. Cerebral organoids exhibit complex biomimetic cortical layering that is lost once samples are dissociated for RNA sequencing and proteomics. Fluorescence microscopy is in principle able to probe the structure of intact organoids on the molecular level at millisecond and nanometer resolution, enabling the dynamic dissection of organoids in response to drugs.

Unfortunately, the potential of a true 4D reconstruction of organoids has been hampered by numerous technical issues. Traditional fluorescence microscopes, like those seen on most highcontent imaging systems, are not capable of clearly imaging 3D samples more than 10 micrometers thick. Computational methods such as 3D deconvolution can assist 3D resolution, but still struggle to deconvolve complex samples like organoids.

Because of these limitations, many organoid imaging projects have traditionally used thin-sectioning methods like cryosectioning to generate thin, near 2D, samples for microscopy. While useful, these methods are labor-intensive, neglect the 3D structure of the organoid, and destroy living tissue. Fortunately, in the last decade a number of methods for the non-destructive observation of 3D tissue have been invented and a number of these approaches are available in commercial devices.

Three dimensional microscopy is available in a number of commercial High-Content Imaging (HCI) systems through confocal microscopy. While many different confocal implementations exist, all these approaches share the same principle – out-of-focus light is excluded from the microscope’s camera by using a pinhole.

Author: Max R. Salick, Eric Lubeck, Adam Riesselman, Ajamete Kaykas,

back to top