Development of a platform for programmable DNA therapeutics
- Reference number
- ITM24-0186
- Project leader
- Zamboni, Margherita
- Start and end dates
- 260101-281231
- Amount granted
- 6 450 689 SEK
- Administrative organization
- Karolinska Institutet
- Research area
- Life Sciences
Summary
Gene therapies are emerging as powerful tools for treating human diseases. Despite initial success in clinical application, several challenges remain to improve safety and efficacy, including the precise regulation of therapeutic genes and cell type-specificity. Indeed, the choice of regulatory elements used in genetic vectors to activate genes relies on limited datasets and screening, thus hindering the design of truly optimized constructs that can achieve highly programmable specificities. My technology aims to address these limitations and bring several conceptual and technological advances, offering a combination of machine learning-based selection of regulatory elements with desired potency and specificity, and their high-throughput experimental validation directly on disease-relevant tissue. The strategic combination of these novel and powerful tools will synergize (1) cutting-edge genomics of human neurological conditions, (2) deep learning models that capture principles of gene regulation and predict enhancer activity, (3) high throughput experimental testing of AI-selected elements, and (4) iterative optimization to generate a robust method for the selection of regulatory elements precisely tailored to specific therapeutic needs. This technology will represent a powerful resource to (pre-)clinical researchers and the biotech sector to guide the design of the next generation of gene therapies achieving unprecedented levels of programmability, safety, and efficiency.
Popular science description
Gene therapies represent attractive precision medicine approaches to treat a wide range of human disorders, being able to specifically target the genetic causes of the disease. These therapies are inspired by the knowledge of how gene expression is regulated in our own cells. Our DNA, present in every cell of the body, represents a long sequence that carries instructions to determine, with exquisite control, which cell will express a certain gene, and to which extent. Gene therapies are designed as synthetic DNA sequences that harbour a therapeutic gene under the control of genetic elements that dictate where and to what level this gene will be expressed. However, a limitation of currently approved DNA therapeutics is that they employ potent genetic regulators to maximize expression of the therapeutic gene. These regulators are active in all (or many) cell types, meaning that the therapeutic gene under their control could be expressed in cells and tissues where it was not intended to be active, potentially leading to adverse effects. Furthermore, the high potency of these constructs might lead to toxicity. We are still lacking a systematic way to predict and better program the activation profile and the specificity of these drugs, to make them safer and more effective. In order to fill this gap, I plan to develop a technology that facilitates the design of optimized genetic constructs with programmable profiles and targeting specificity. The technology will be developed using large genomics data of human neurological conditions, and it will take advantage of powerful machine learning models to understand and predict the regulatory activity of the genetic elements under investigation. It will, furthermore, encompass a platform to perform highly parallelized validation of these elements directly on clinical tissue. Finally, the technology will be deployed on human samples of traumatic brain injury and neurodegenerative disease, to aid the discovery of clinically-relevant genetic elements that could be used for new gene therapy approaches. Overall, this project will open the door to the creation of the next generation of programmable DNA therapeutics, optimized for their safety and efficacy.