CRISPR-Cas9 in Disease Modeling

As cell-based disease models continue to replace test animals in drug development, researchers require cell lines that represent disease states more accurately and faithfully than lines derived solely from primary healthy cells, progenitor/stem cells, or biopsies of diseased tissue. The answer for achieving this final and necessary step in disease modeling is gene editing, specifically through CRISPR-Cas9.

The emergence of 3D cell cultures (spheroids, organoids, organ-chips, etc.) has changed drug discovery by creating more lifelike representations of disease states, but even these models require cells carrying specific genetic characteristics. Harvesting primary cells is limited by their availability in living patients (not to mention the pain and cost involved in their acquisition) and the inherent heterogeneity when obtained through biopsy. The differentiation of induced pluripotent stem cells (iPSCs) generates organ- or tissue-specific cells, but these approaches take the cell-based model idea only so far.

Enter the editors

Gene editing was for decades at the top of the wish list for molecular biologists, but unrealized until the discovery of nucleases—specifically zinc finger nuclease (ZFN) in 1985—which made gene deletions and insertions practical and accessible. This was followed by the introduction of CRISPR-Cas9 (1987), and later transcription activator-like effector nucleases (TALENs; 2010).

CRISPR works by snipping foreign DNA into fragments approximately 20 base pairs in length, and pasting them into continuous sequences (CRISPR arrays) —essentially a cut-and-paste operation. For a while CRISPR competed in the highly competitive gene-editing market with ZFN and TALENs, but CRISPR eventually came to dominate.

CRISPR consists of two parts. The Cas9 enzyme that creates a break in the target gene, and the guide ribonucleic acid (RNA) sequence, which the enzyme deploys to locate the precise region where editing will occur.

CRISPR (and similar gene manipulations) edits cells or organisms at the genome level, and is capable of editing multiple genes. If the target sequences are selected carefully, these changes may be reflected in the phenotype as well.

CRISPR in 3D

Perhaps the greatest advantage of CRISPR in modeling diseases is its combination with other established models, especially with three-dimensional (3D) spheroids, organoids, tissue- or organ-chips, etc. 3D cultures provide not just the genotype or phenotype under investigation, but a reasonable facsimile of the cells’ in vivo biological niche.

The main issue with organoids, in particular, is the availability of cells that represent the pathology under study. Primary cells are difficult to harvest, especially from living organisms, and their ability to replicate is limited. Organoids derived from induced pluripotent stem cells can theoretically differentiate into any cell type, e.g., fibroblast, neuron, cardiomyocyte, but these typically carry wild-type (i.e., “healthy”) genes.

CRISPR effects genomic changes that more completely reflect the disease state through the introduction of one or more gene edits, resulting in novel cells that more faithfully reflect the disease phenotype.

CRISPR models in action

Patient-derived cells have been used for decades as tumor models, but conventional test cultures derived from prepared biopsy samples lack the cellular diversity, spatial organization, and the microenvironment that make cancerous tissues unique. Hence the 3D culture idea. The advantage of applying CRISPR-Cas9 editing to these systems is the potential to have not just the cells and the niche, but to incorporate the molecular changes defining a disease state.

The application of genome-editing broadens the type and level of changes that may be introduced into these cells, while reducing the risk of “stray” phenotypes arising or off-target effects.

Stem cells are the starting points for a great deal of cell-based assay work, but differentiation protocols take this idea only so far. The creation of cell lines reflecting not just the gross phenotype (e.g., “cardiomyocyte”), but the specific mutation, insertion, or deletion characteristic of specific diseases, requires gene editing. Fortunately, the application of most gene-editing methods, including CRISPR, before differentiation does not usually affect the cells capacity for either differentiation or expansion.

For example, brain organoids are the basis of physiologically relevant models for studying brain development, gene expression, and cellular maturation. That’s the 35,000-foot view. To gain insights into more granular processes, e.g., temporal transcriptional signature, dynamic cell structure, and electrophysiological responses, researchers have applied CRISPR-Cas9 editing to produce more nuanced models through advanced maturation.

A group at Capital Medical University, Beijing, generated an hERG-deficient cardiomyocyte model that mimics long-QT syndrome, a hereditary heart arrhythmia. Researchers first used CRISPR to remove KCNH2, a gene that encodes for a potassium channel subunit involved in long-QT. They found the knockout did not affect the stem cells’ pluripotency or differentiation capacity. The hERG-deficient cells, which showed irregular rhythm, were subsequently used to test potential treatments.

An Australian research team has used CRISPR to create novel breast cancer models from breast progenitor cells. After generating organoids through controlled differentiation, they applied CRISPR to knock out the relevant tumor suppressor genes (P53, PTEN, RB1, NF1). Gene-modified organoids achieved the capacity for long-term culturing and formed estrogen-receptor positive lumina. Moreover, they responded appropriately to endocrine or drug therapy which, according to the authors, supports “the potential utility of this model to enhance our understanding of the molecular events that culminate in specific subtypes of breast cancer.”

Conclusion

CRISPR/Cas9 has several advantages over ZFN- and TALEN-based editing methods, including lower cytotoxicity and more-efficient targeting. CRISPR is also capable of acting at multiple locations in the genome.

However, no gene-editing method is without drawbacks or caveats. Generating CRISPR-edited human pluripotent stem cells is a multi-step process that includes transfection, isolation, screening, and most importantly in-depth molecular validation.

Of greater concern, the presence of a protospacer-associated motif (PAM), which occurs once in roughly eight base pairs, may prevent targeting of the CRISPR nuclease. Another hurdle is the relatively high frequency of off-target cleavage. CRISPR variants have been developed to counter these deficiencies. For example, Cas9 nickase (Cas9n) generates a single-strand DNA break instead of wild-type CRISPR’s double-stranded break. Cas9n has been shown to reduce off-target effects while retaining the targeting efficiency of the original CRISPR-Cas9.