Research into neurodegenerative conditions can be hampered by difficulty accessing scalable, translatable cellbased disease models. Amanda Turner charts the hunt for solutions.
Huntington’s disease is a devastating inherited neurodegenerative condition that affects approximately five to 10 out of every 100,000 people [1]. In the general population, the huntingtin gene contains a repeating cytosine-adenine-guanine (CAG) DNA sequence with between six and 35 copies. Individuals with over 40 CAG repeats in their huntingtin gene will develop Huntington’s disease in their lifetime. The mutation negatively impacts a variety of neuronal cells within the brain over time, causing irreversible damage and ultimately premature death.
Unfortunately, there is no known cure for this disease, and to date, the underlying mechanisms remain poorly understood. Consequently, there are limited treatment options for patients, with existing therapies focusing on managing symptoms rather than treating the underlying cause. Finding new, safe, and effective therapeutics to treat neurodegenerative diseases like Huntington’s disease relies on the availability of cell models that accurately mimic the condition for use in preclinical validation and testing.
Huntington’s disease researchers are currently limited by a lack of translatable, easily scalable models that provide direct physiological relevance to the condition. Any technology that can produce accurate cell-based disease models will greatly benefit the community of scientists working on neurodegenerative diseases, and ultimately patients.
The challenge with existing disease models
At present, scientists utilise disease models from a variety of sources to study and identify new therapies for neurodegenerative disease. The models include animals, primary cell models, or iPSC-derived neurons. All of these offer some level of translatability. However, in practice, each has limitations.
Taking Huntington’s disease as an example, existing animal models have been challenging to scale and translate into the clinic for therapeutic discovery. Each model has a differing genetic background and a differing number of CAG repeats, meaning their disease phenotype can vary considerably. Results from one model may not be repeatable in another, requiring significant cross-examination for reliable conclusions to be drawn. Similar issues affect research with cells derived directly from human or animal tissue. Each new batch of cells needs to be fully validated due to variability in donor and organism of origin, meaning their utilisation is limited to low throughput workflows. These cells also lack genetically matched controls, making it tough to attribute observed effects back to the disease-causing mutation. Specifically, neurons donated by individuals who had Huntington’s disease are in limited supply, meaning scientists often need to seek out alternative options for in vitro disease modelling [1].
New cell models will help enhance fundamental research into such devastating conditions, and improve the successful translation of new potential therapies into the clinic
Human induced pluripotent stem cell (iPSC) derived cells have emerged as a promising tool for disease modelling. In theory, such models should be both consistent and scalable, as every cell can be generated from the same stem cell donor background using the same methodology, and the donor can be cultured ad infinitum. In practice though, the generation of disease models from iPSCs still suffers inconsistency and batch-to-batch heterogeneity.
Scientists commonly derive Huntington’s disease models from patient-derived iPSCs using directed differentiation. Cocktails of molecules responsible for driving cell fate are added to the stem cell culture at specific time points over a period of months to mimic neuron development in vivo. Not only are directed differentiation protocols laborious, complex and time-consuming, but the requirement for scientists to routinely perform multiple complicated steps can also increase the burden of user-derived variance.
A 2021 study compared two methods from two different labs for the directed differentiation of a patientderived stem cell line into neuronal cells that model Huntington’s disease [2]. It was shown that even though both methods used identical chemical cocktails for cell differentiation, slight variations in the protocols and culture conditions led to neurons that displayed vastly different phenotypes. Additionally, when assessing the proportion of desired cells versus other cell types in the final population, batch-tobatch variation swung wildly, from five per cent to 64 per cent. This lack of consistency batch-tobatch and protocol-to-protocol ultimately means the directed differentiation of patient iPSCs generates unreliable models for the study of Huntington’s disease. Additionally, as each new batch of cells would need thoroughly validating before translatable results can be generated, throughput remains a limitation.
Hunting for a consistent and scalable model
‘Precision reprogramming’ is a recent technology for the generation of batch-to-batch consistent cell models developed by Dr Mark Kotter’s lab at the University of Cambridge [3]. This method uses an inducible gene expression system inserted into ‘safe harbour sites’ within the genome of human stem cells. Safe harbour sites are conserved regions of the genome that are immune to gene silencing. By controlling the expression of cell fate defining transcription factors with this system, iPSC populations can be reprogrammed into the desired cell type in a consistent and predictable manner. Kotter’s lab named the technology “optimised inducible overexpression”, or opti-ox. Since its development, Kotter has spun opti-ox out into two companies: Meatable, focused on lab-grown meat, and bit.bio, focused on the precision reprogramming of stem cells.
bit.bio has since combined opti-ox reprogrammed neurons with CRISPR/Cas9 gene editing to generate neurodegenerative disease-specific models [4]. Recently, their first precision reprogrammed human disease model for Huntington’s disease in a glutamatergic neuron background was launched. An abnormal 50 CAG repeat expansion has been introduced into the cell’s huntingtin gene, accurately reflecting the disease’s genetics (Fig 1, see above, full caption at end of post).
Precision reprogrammed iPSC-derived disease models come with a critical advantage: scientists can pair the disease model with a precision reprogrammed wild-type cell to produce a genetically matched isogenic pair. Comparing the Huntington's disease model to the isogenic control allows scientists to identify and investigate the effects of the disease mutation directly, aiming to provide more accurate, reproducible, and translatable data (Fig 2, see below, full caption at end of post).
As every cell in a population of precision reprogrammed stem cells has the same identity, opti-ox overcomes the fundamental challenges of batch-to-batch variability and cell availability. Additionally, the ability to generate large volumes of desired cells in a matter of days, in a consistent and predictable manner makes precision reprogrammed disease models applicable to higher throughput screening methods [3]. It will be exciting to see how this new type of disease model accelerates therapeutic development for neurodegenerative diseases like Huntington’s.
Positive impact
Access to a disease model with batch-to-batch consistency, scalability to high throughput screening, a disease-relevant genotype and a genetically matched isogenic control can significantly enhance the disease modelling process. The availability of such models can help accelerate the identification of therapeutic targets and molecules, while helping to improve our understanding of neurodegenerative diseases like Huntington’s disease.
Ultimately, these new cell models will help enhance fundamental research into such devastating conditions, and improve the successful translation of new potential therapies into the clinic, positively impacting patients and their families.
Amanda Turner is Senior Product Manager at bit.bio
References:
1 Mochly-Rosen D, et al. The challenge in translating basic research discoveries to treatment of Huntington disease. Rare Dis 2:e28637 (2014). doi: 10.4161/rdis.28637
2 Le Cann K, et al. The difficulty to model Huntington’s disease in vitro using striatal medium spiny neurons differentiated from human-induced pluripotent stem cells. Sci Rep 11, 6934 (2021). https://doi.org/10.1038/s41598-021-85656-x
3 Pawlowski M, et al. Inducible and Deterministic Forward Programming of Human Pluripotent Stem Cells into Neurons, Skeletal Myocytes, and Oligodendrocytes. Stem Cell Rep 8(4): 803–812 (2017) 4 Charles River. A scalable, reproducible platform for reprogrammed human cells. Nature https://www.nature.com/articles/d42473-021- 00584-8 (2022)
Captions:
Gel electrophoresis: Fig 1. (A) Successful on-target integration into one huntingtin (HTT) allele confirmed by gel electrophoresis. Genotyping primers flanking the endogenous HTT CAG repeat expansion region produce a band at approximately 320bp, by PCR, in both isogenic control (ioGlutamatergic Neurons) and disease model (ioGlutamatergic Neurons HTT50CAG/WT. PCR fragments at 395bp detect on-target gene editing and introduction of a 50 CAG repeat expansion in ioGlutamatergic Neurons HTT50CAG/WT only. (B) PCR amplification of the donor vector backbone, used to integrate the 50 CAG repeat expansion at the wild-type (WT) HTT locus, verified that no off-target random insertion had taken place in the samples from ioGlutamatergic Neurons HTT50CAG/WT
Immunocytochemistry/Immunofluorescence: Fig 2. Immunofluorescent staining of cells post-revival demonstrates similar homogenous expression of panneuronal proteins (MAP2 and TUBB3) and glutamatergic neuron-specific transporter (VGLUT2) in bit.bio’s ioGlutamatergic Neurons HTT50CAG/WT and in the isogenic control, by Day 11 (100X)
