Combining Machine Learning and Multiplexed, In Situ Profiling to Engineer Cell Type and Behavioral Specificity

Area of Science:

• Neuroscience
• Genetics
• Bioengineering

Background:

• Precise control of neural circuits is crucial for understanding and treating neurological disorders.
• Discovering cell-specific enhancers for gene expression is challenging due to low in vivo success rates and species-specific activity.
• Current methods lack spatial detail and struggle with multiplexing viral vectors like adeno-associated viruses (AAVs).

Purpose of the Study:

• To accelerate the discovery of cell-targeting enhancers for the dorsal spinal cord, a key region for pain and itch processing.
• To develop an integrated platform combining machine learning, modular AAV assembly, and in situ screening for enhancer discovery.
• To enable spatially resolved, multiplexed in vivo screening of enhancers.

Main Methods:

• Developed ESCargoT (Engineered Specificity of Cargo Transcription), an end-to-end platform for enhancer discovery.
• Utilized cross-species chromatin accessibility data to train machine learning models for predicting enhancer activity in specific cell types.
• Created a Spatial Parallel Reporter Assay (SPRA) integrating Golden-Gate assembly with multiplexed in situ screening for parallel profiling of an enhancer-AAV library.

Main Results:

• Successfully identified and validated enhancers targeting excitatory dorsal horn neurons, leading to the reversal of mechanical allodynia.
• Demonstrated the ability to screen a library of 27 enhancers simultaneously in vivo, targeting diverse cell types including oligodendrocytes and dorsal horn neuron subtypes.
• Validated enhancers targeting Exc-LMO3 and Exc-SKOR2 neurons, with one shown to block chemical itch sensation.
• Showcased cross-species applicability, with macaque-derived enhancers functional in mice.

Conclusions:

• The ESCargoT platform significantly accelerates the discovery of cell-specific enhancers for neural circuit manipulation.
• Spatially resolved, multiplexed in vivo screening is effective for identifying functional enhancers.
• This approach facilitates the development of novel cell-targeting tools and gene therapies for pain and itch disorders.