MERFISH 2.0, an ultra-sensitive single-cell spatial transcriptomics imaging chemistry across diverse tissue types

The authors developed MERFISH 2.0, an optimized spatial transcriptomics chemistry that significantly enhances transcript detection sensitivity (up to ~8-fold) in degraded and archival tissues like FFPE samples compared to its predecessor, thereby enabling robust, scalable, and high-resolution spatial analysis across diverse clinically relevant specimens.

The Gist

Imagine you are trying to read a library of books to understand how a city works. In this library, the "books" are the instructions inside your cells (RNA), and the "city" is your body's tissue.

For a long time, scientists have had a powerful tool called MERFISH to read these books. It's like a high-tech scanner that can read thousands of specific words in a book while keeping the book open on the shelf, so you know exactly where in the city that word was spoken.

However, there was a big problem: Old or damaged books.

In the real world, many of the most valuable medical records are kept in archives. These are tissue samples taken from patients years ago, preserved in wax (a process called FFPE). Over time, the "ink" in these books fades, the pages get stuck together, and the text breaks into tiny fragments. The old scanner (MERFISH 1.0) struggled to read these damaged books. It would miss a lot of words, leaving scientists with a blurry, incomplete picture of the disease.

Enter MERFISH 2.0: The Super-Scanner.

The scientists at Vizgen developed a new version of this technology, MERFISH 2.0, which acts like a "super-reader" specifically designed to handle these messy, damaged archives.

Here is how they did it, using simple analogies:

1. The Sticky Tape (Better Anchoring)

• The Problem: In the old method, when they tried to scan the tissue, some of the broken RNA fragments would fall off the page and get lost in the wash.
• The Fix: MERFISH 2.0 uses a special "sticky tape" (optimized anchoring) that grabs onto even the tiniest, most broken fragments of RNA and holds them tight to the page. This ensures that even the most damaged samples don't lose their data.

2. The Magnifying Glass (Better Probes)

• The Problem: The "ink" in old samples is cross-linked (stuck together), making it hard for the scanner's probes to find the words they are looking for.
• The Fix: They redesigned the scanner's "fingers" (probes). Instead of just trying to grab the word, these new fingers have a special shape that can wiggle through the sticky, cross-linked mess and grab the target much more efficiently.

3. The Spotlight (Signal Enhancement)

• The Problem: In old samples, the signal (the light from the words) is very dim, like trying to read a candle in a dark room.
• The Fix: MERFISH 2.0 introduces a "signal booster." It's like adding a magnifying glass that focuses light onto the words. It makes the faint signals bright and clear without making the background noise louder. This allows the scanner to see words that were previously invisible.

What Did They Find?

The researchers tested this new scanner on all kinds of tissues: fresh mouse brains, human brains, and even old, archived cancer samples.

• The "8-Fold" Boost: In the worst-quality, oldest samples, MERFISH 2.0 found 8 times more words than the old version. It's like going from reading a book with 100 words to reading a book with 800 words.
• Finding Hidden Characters: Because the scanner is so much more sensitive, it found cell types that were previously invisible. For example, in old breast cancer samples, they found many more immune cells (the body's "police") hiding in the tumor. The old scanner missed them because they were too quiet; the new scanner heard them clearly.
• Better Maps: With more data, the scientists could draw a much more detailed map of the tumor. They could see exactly how the cancer cells were talking to the immune cells, which is crucial for figuring out how to treat the disease.

Why Does This Matter?

Think of the world's medical archives as a giant warehouse of old, dusty files. For years, scientists could only read the fresh, clean files. Now, with MERFISH 2.0, they can finally open the dusty, damaged files and read them clearly.

This means:

1. Better Research: Scientists can use decades of old patient data to study diseases like cancer and Alzheimer's with a level of detail that was impossible before.
2. Personalized Medicine: By understanding the exact "neighborhood" of cells in a patient's tumor, doctors might be able to choose better treatments.
3. No More Waste: We don't have to throw away old samples; we can finally get the full story out of them.

In short, MERFISH 2.0 is a game-changer that turns blurry, incomplete medical records into high-definition movies, helping us understand the complex city of our bodies better than ever before.

Technical Summary

1. Problem Statement

Spatial transcriptomics is critical for understanding tissue architecture and cellular heterogeneity, but current imaging-based methods face significant limitations when applied to archival and clinically relevant specimens, particularly Formalin-Fixed, Paraffin-Embedded (FFPE) tissues.

• The Challenge: FFPE tissues suffer from RNA degradation (fragmentation) and extensive chemical crosslinking caused by formalin fixation.
• The Consequence: In the original MERFISH chemistry (MERFISH 1.0), these factors reduce probe accessibility, limit binding sites on fragmented transcripts, and increase background fluorescence. This leads to poor transcript detection sensitivity, high cell dropout rates, and compromised downstream single-cell analyses in low-quality samples.
• The Gap: While MERFISH 1.0 performs well on high-quality fresh-frozen samples, its utility is restricted in the vast repositories of archival FFPE samples essential for clinical research and translational medicine.

2. Methodology: MERFISH 2.0 Chemistry and Workflow

The authors developed MERFISH 2.0, an optimized imaging chemistry and unified sample preparation workflow designed specifically to handle fragmented and highly crosslinked RNA.

Key Technical Innovations:

1. Optimized RNA Anchoring: The protocol was modified to improve the immobilization of fragmented RNA transcripts within the hydrogel during sample preparation. This ensures that short RNA fragments are retained rather than lost during the clearing process.
2. Redesigned Probe Structure: The encoding probes were redesigned. Unlike MERFISH 1.0, where readout bits are directly conjugated to the target, MERFISH 2.0 uses an overhang structure that binds to the target region and subsequently hybridizes with enhancer probes. This design facilitates more efficient hybridization on fragmented targets.
3. Controlled Signal Enhancement: A new "enhancer" step was introduced to amplify imaging signals in a controlled manner, significantly boosting the signal-to-noise ratio without compromising quantitative accuracy.
4. Unified Workflow: MERFISH 2.0 established a single, unified workflow for both fresh-frozen and FFPE samples. In this workflow, samples are first gel-embedded and cleared, followed by encoding probe hybridization. This contrasts with MERFISH 1.0, which required distinct, often suboptimal, workflows for frozen vs. FFPE samples.

Experimental Design:

• Samples: The study benchmarked MERFISH 2.0 against MERFISH 1.0 across diverse tissue types:
  • Mouse: Fresh-frozen brain; Fixed-frozen liver, heart, kidney, lung, intestine, spleen, and brain.
  • Human: Fresh-frozen brain, archival fresh-frozen brain, and archival FFPE breast cancer and lung cancer tissue microarrays (TMAs).
• Gene Panels: 815-plex MERFISH panels were used, including specific panels for neurobiology, immunology, and breast cancer.
• Analysis: Data was processed using the MERSCOPE pipeline, with cell segmentation via Cellpose and downstream analysis (clustering, correlation, spatial enrichment) using Scanpy and Squidpy.

3. Key Contributions

• Development of MERFISH 2.0: A novel chemistry that overcomes the sensitivity barriers of FFPE and degraded RNA samples.
• Unified Protocol: A streamlined, single workflow applicable to both frozen and FFPE tissues, reducing technical variability.
• Comprehensive Benchmarking: A rigorous head-to-head comparison across multiple species, tissue types, and preservation methods, establishing quantitative metrics for sensitivity and concordance.

4. Key Results

A. Sensitivity Gains:

• High-Quality Samples (Fresh-Frozen Mouse Brain): MERFISH 2.0 showed a ~2-fold increase in transcript detection per 100 µm² and per cell compared to MERFISH 1.0, with reduced biological variability.
• Archival/Decayed Samples (Human Brain & FFPE): The improvements were most dramatic in low-quality samples.
  • Archival Fresh-Frozen Human Brain: ~5-fold increase in transcript density and ~2.5-fold increase in transcripts per cell.
  • FFPE Breast Cancer: Up to ~8-fold increase in transcript detection in low-quality cores.
  • General Trend: Across all FFPE and fixed-frozen tissues, sensitivity increased by 2- to 8-fold depending on sample quality.

B. Quantitative Concordance:

• Despite the massive gain in sensitivity, MERFISH 2.0 maintained high quantitative fidelity.
• Correlation coefficients between MERFISH 1.0 and 2.0 data were consistently high (Pearson r ≥ 0.8, often >0.9), indicating that the increased signal reflects true biological abundance rather than noise.
• MERFISH 2.0 data showed stronger correlation with bulk RNA-seq and single-cell RNA-seq (scRNA-seq) reference data than MERFISH 1.0.

C. Biological Insights & Downstream Analysis:

• Reduced Cell Dropout: In FFPE breast cancer samples, the number of cells passing quality control (QC) nearly doubled (from ~367k to ~654k), leading to a denser, more continuous spatial map.
• Improved Cell Type Resolution:
  • Brain: MERFISH 2.0 enabled the identification of rare cell types (e.g., MGE interneurons) and clearer delineation of cortical layers that were obscured in MERFISH 1.0 data.
  • Tumor Microenvironment: MERFISH 2.0 detected significantly more immune cells (T cells, B cells, myeloid cells) and identified a distinct population of cycling myeloid cells missed by MERFISH 1.0.
• Enhanced Interaction Analysis:
  • Cell-Cell Interactions: The higher cell retention allowed for robust neighborhood analysis, revealing clear spatial enrichments (e.g., B-T cell associations, endothelial-perivascular interactions) that were statistically weak or invisible in MERFISH 1.0.
  • Gene-Gene Correlation: MERFISH 2.0 strengthened the detection of co-regulated gene modules (e.g., STAT2/JUN/STAT6, GATA3/AKT1), facilitating better inference of regulatory pathways.

5. Significance

• Unlocking Clinical Archives: MERFISH 2.0 transforms the utility of the vast global repository of FFPE tissue blocks, enabling high-resolution spatial transcriptomics on samples that were previously considered too degraded for analysis.
• Translational Impact: By improving sensitivity in low-quality samples, the technology enables more accurate immunoprofiling, biomarker discovery, and therapeutic target identification in clinical cohorts.
• Robustness and Reproducibility: The unified workflow and reduced variability make MERFISH 2.0 suitable for large-scale, multi-site cohort studies where sample quality is often heterogeneous.
• Future of Spatial Biology: This work demonstrates that optimizing chemistry for RNA fragmentation and crosslinking is a critical step toward making spatial transcriptomics a standard tool in pathology and precision medicine.