BALTIMORE – Researchers from Massachusetts General Hospital (MGH) and their collaborators have developed a new method that employs microscopy to enable spatially resolved single-cell sequencing.
Named Image-seq, the method, which was described in a Nature Methods paper published in November 2022, deploys in vivo or in situ microscopy to help isolate intact and viable cells from specific locations within a tissue for downstream single-cell sequencing.
“We had this unique idea of combining spatially resolved transcriptional information, or even multiomics information down the line, with multiphoton microscopy and in vivo imaging,” said Christa Haase, a researcher at MGH and the first author of the study.
The core element of Image-seq is a multiphoton microscope, which has one optical path for imaging and an additional laser for tissue ablation. As the ablation laser performs precise microdissections on the sample tissue, a micropipette can aspirate the targeted cells under image guidance.
Once isolated, the cells can go into existing single-cell library preparation workflows, followed by sequencing and data analysis. “With our technology, since you have these live cells, you can make use of all of the advances that are happening in the entire field,” Haase noted.
According to Haase, one of the unique advantages of Image-seq is its ability to tackle sample types, such as bone marrow, that have been historically challenging for most spatial analysis workflows. “Most spatial transcriptomics [methods] require generating some sort of thin tissue slice,” she explained. “One of the challenges with bone marrow and bone is that these are tissues that are very difficult to section.”
In addition, Haase said the method can help enrich rare cell types that are not easily detectable by conventional spatial transcriptomics workflows. Furthermore, Image-seq can provide researchers with a window into the temporal dynamics of cells by tracking their activities over time under in vivo imaging, she said.
In their study, the MGH researchers employed Image-seq to study leukemia biology in bone marrow, the primary site for hematopoiesis.
Specifically, they combined Image-seq with high-throughput, droplet-based sequencing using the 10x Genomics Chromium platform to study bone marrow hematopoietic cells, as well as with the SMART-Seq v4 protocol to profile rare acute myeloid leukemia (AML) cells and bone marrow stromal cells. The authors also tracked AML progression under the microscope by observing the spatial heterogeneity of the cells.
With the help of Image-seq, they identified dipeptidyl peptidase 4 (DPP4) as a key upregulated gene in AML cells in the more proliferative bone marrow compartments. “Strikingly, DPP4 was not expressed on the same cells cultured in vitro, suggesting that DPP4 was specifically activated in vivo and was correlated with disease progression,” the authors reported.
“It's a pretty neat and powerful approach for single-cell analysis,” said Junyue Cao, head of the single-cell genomics and population dynamics laboratory at Rockefeller University. “I think the unique part of this strategy is that it is developing a novel cell isolation approach and coupling that with imaging.”
Echoing Haase’s point, Cao said one strength of Image-seq appears to be its capability to isolate and help study critical cell types that are often too rare to be captured by global spatial transcriptomics profiling. Additionally, Cao praised the method's ability to capture the temporal dynamics of the cells while being compatible with conventional single-cell genomic workflows.
However, despite the method’s promises, Cao thinks Image-seq might still be “relatively challenging for normal labs” to carry out. “It's not very easy for an average lab to get access to the equipment, especially the specialized lasers,” he pointed out, adding that the method may also require some special training for researchers to execute it in their own lab.
Depending on the type of experiment, Haase said, the team typically harvests between 500 and a few thousand cells for each Image-seq procedure, which takes about 15 minutes per sample.
Although the group has achieved single-cell resolution for rare leukemia cells, Haase said, in that case, the throughput was compromised. “Because we're aspirating out cells in a specific spatial location, the more cells we take out, the more resolution we lose,” she explained, noting the current trade-off between throughput and resolution for Image-seq.
One future direction for the method is to continue improving its resolution while increasing its throughput, she said. Additionally, since the current cell isolation procedure for Image-seq is semi-manual, another possible further direction is to automate the workflow.
The team has not yet tried Image-seq on formalin-fixed, paraffin-embedded (FFPE) tissues, but “there's no reason why it shouldn't be” working on these samples, Haase said.
When it comes to cost, Haase said if a lab is equipped with a confocal microscope that meets Image-seq’s specifications, the method should be “as expensive as any other single-cell sequencing experiments.” Meanwhile, some labs may still need to purchase an external laser as an addition to the microscope to carry out some of the functions of Image-seq.
The MGH team has applied for patents for Image-seq and is considering its commercial development. “We're thinking of commercializing [the] technology as an add-on to a multiphoton microscope,” Haase said. “I think options that are being explored are existing startup companies or larger companies that may already have a similar type of framework that this could be integrated into.”
Moving forward, Haase said her team is planning to apply Image-seq to help investigate diseases and biological questions related to the bone marrow, as well as other tissues that remain difficult to study using other single-cell analysis tools.
“Our ultimate goal is to use insights that we can gain with this technology to develop new treatments for blood disorders,” she said.