Abstract
Expansion-Assisted Iterative Fluorescence In Situ Hybridization (EASI-FISH) is a powerful technique that integrates expansion microscopy with multiplexed FISH to investigate gene expression in thick tissue samples. We describe an updated fly EASI-FISH protocol suitable for high throughput experimentation. Up to six brains or two CNS can be embedded in a single gel and several gels can be processed in parallel. We use a new gel recipe that improves gel robustness and permits at least 10 rounds of hybridization. Using the GAL4-UAS system for co-detection of green fluorescent protein (GFP), gene expression can be visualized in cell types of interest. Due to the high resolution and sensitivity of the method, single transcripts can be detected, and spot counting can be used to quantify expression levels.
Introduction
Mapping gene expression within the complex architecture of the brain is essential for understanding neural circuits. However, researchers face two primary obstacles: the inherent thickness and opacity of intact brain tissue, which scatters light and limits resolution, and the "color barrier" of standard fluorescence microscopy, which typically allows for only 3–4 targets to be imaged simultaneously.
Expansion-Assisted Iterative Fluorescence In Situ Hybridization (EASI-FISH) was developed to overcome these hurdles. By physically expanding the tissue and using iterative rounds of probing, it enables the mapping of dozens of mRNA targets. This paper introduces EASI-FISH v2, a refined protocol optimized for high-throughput, robust, and sensitive whole-brain molecular mapping in Drosophila.
The EASI-FISH v2 Workflow
The v2 protocol improves upon its predecessor through several key technical advancements designed to make spatial transcriptomics more accessible for large-scale experiments.
Expansion and Clearing
The process begins by embedding the Drosophila brain in a specialized hydrogel. RNA and proteins are chemically anchored to the gel matrix using Melphalan-X and Acryloyl-X. Following enzymatic digestion of structural proteins, the gel is expanded (typically 2x to 4x in linear dimension), which physically separates molecules beyond the diffraction limit of light and renders the tissue transparent for high-resolution imaging.
Iterative Signal Amplification
Instead of traditional FISH, the protocol employs Hybridization Chain Reaction (HCR). HCR triggers the polymerization of fluorescent hairpins at each mRNA target site, providing significant signal amplification. After imaging, probes are stripped using DNase, allowing for subsequent rounds of hybridization. The new TREx-1000 gel recipe ensures the sample remains mechanically stable through at least 10 such iterations without warping or signal degradation.
High-Throughput Processing
Efficiency is a core focus of version 2. A new mounting technique allows up to six brains to be embedded in a single gel. By processing multiple gels in parallel, researchers can significantly scale up their experiments, moving toward "atlas-scale" mapping projects.
Single-Molecule Sensitivity and Protein Co-detection
A major advantage of EASI-FISH v2 is its ability to preserve endogenous protein signals, such as GFP. This allows researchers to utilize the GAL4-UAS system to identify specific cell types or circuits while simultaneously measuring gene expression.
Because expansion physically increases the distance between mRNA molecules, "spot counting" becomes highly accurate. This transforms the analysis from qualitative (on/off) to quantitative, providing discrete transcript counts per cell while maintaining the spatial context of the entire brain architecture. Unlike single-cell sequencing, which loses the physical location of cells, EASI-FISH v2 keeps the entire brain architecture intact, allowing researchers to see how gene expression changes across specific brain regions or along neural projections.
Results and Future Directions
The protocol demonstrates the capacity for over 10 rounds of hybridization with single-transcript resolution. While highly effective, the authors note that certain far-red dyes (like 647) are prone to photobleaching in the gel, recommending alternatives like CF633 or Janelia Fluor 669 for better stability.
The primary remaining challenge is the computational bottleneck. Generating terabytes of data from multiple rounds of whole-brain imaging requires robust automated pipelines. Tools like Cellpose-SAM for cellular segmentation and fishspot for automated transcript counting are currently being integrated to streamline the transition from raw imagery to biological insight.
Summary updated