Blog
May 5, 2023 / Oncology

Connecting single cells to spatial context, researchers to revolutions, at AACR 2023

Olivia Habern

AACR 2023 brought together a global community of cancer researchers with a shared passion to advance our knowledge of cancer biology and mechanisms of therapeutic response and resistance. With this knowledge, the community hopes to produce more effective treatment strategies and influence improved long-term outcomes for patients. We were inspired and honored to see the work that so many scientists have been doing to achieve these goals, aided by single cell and spatial technology.

Connecting these worlds of insights—highly dimensional single cell data with crucial spatial context—represented a major theme of the conference. One question that comes to mind is: what is the potential of this approach to ultimately unravel intratumoral heterogeneity and dissect the immune and tumor microenvironment? 10x’er Abbey Cutchin, Associate Director of Segment Marketing and conference attendee, shared her excitement for what cancer researchers can now do with high-throughput, high-resolution tools that span the single cell and spatial planes of biology:

“It feels like we're on the cusp of a watershed moment where new technology is going to lead to a new wave of discovery by enabling and unlocking questions that were just not possible to ask and answer before. And then at scale, too. I'm most excited for the novel insights that we'll be able to see as these technologies, like all of the in situ and spatial approaches, start to mature and single cell analysis becomes more ubiquitous and used at greater scale. We should be able to gain statistically powerful insights and leverage single cell technology beyond just a discovery tool, but actually deploy it across larger cohorts.”

In this post, we highlight some of the innovative experiments using single cell and spatial approaches that were showcased at AACR 2023, and conference conversations about these new technological capabilities. Keep reading to explore:

  • Efforts to forecast the trajectory of precancerous lesions in pancreatic cancer (Jump to section)
  • New methods to resolve diverse immune and cancer cell states and their spatial localization in archived clinical breast cancer samples (Read now)
  • What cancer researchers are saying about the potential applications for Xenium In Situ analysis (Skip to section)
  • Possibilities for retrospective studies using combined single cell and spatial analysis in FFPE tissue with Chromium Single Cell Gene Expression Flex and Visium CytAssist (Jump ahead)

Forecasting the precancer microenvironment with high-dimensional single cell and spatial analysis

Beyond cell-type identification, what’s the ultimate value of the data that cancer researchers can get from single cell analysis of heterogeneous tumor samples? According to Elana Fertig, PhD, Associate Professor of Oncology and Director of the Division in Quantitative Sciences at Johns Hopkins University Sidney Kimmel Comprehensive Cancer Center, high-dimensional single cell data from a tumor lesion can reveal transcriptional patterns dictating biological variation in the sample and the genes behind that variation. This information offers clues that can help scientists understand the cell state transitions between normal cells and cancer cells present in the sample. These patterns then become the lens through which to evaluate other transcriptional datasets, including datasets taken at different stages of tumor progression, and provide a path to forecast carcinogenesis and therapeutic interception. Datasets derived from spatial assays can also be explored for these same patterns.

Applying these principles, Dr. Fertig and her team used single cell RNA-seq to establish a standardized cellular atlas of pancreatic ductal adenocarcinoma. Using a computational method called CoGAPS, they teased out two key patterns associated with cellular proliferation and inflammation in normal tumor-adjacent epithelial cells, which they conjectured might represent precursor cells to malignancy. They next sought to relate these two patterns across stages of pancreatic carcinogenesis in a spatially resolved manner: aided by a computational method called ProjectR, they used Visium to map the proliferation and inflammatory patterns in FFPE tissue sections taken from normal pancreatic tissue and low-grade or high-grade precancer tissue. The Visium data revealed a colocalization of the two patterns, but an inverse relationship between their expression signatures, where the proliferation pattern increased over the course of tumor progression while the inflammatory pattern decreased.

To better resolve the possible regulatory mechanisms behind these observations, Dr. Fertig’s group used Xenium In Situ analysis to achieve true single cell–resolution spatial transcriptomic profiling of the precancer microenvironment. Using a 400-gene cancer panel and a customized panel built to contain gene expression signatures they’d identified in their patterns, the researchers observed the same trade off in proliferation and inflammation patterns surrounding the pancreatic duct. They also observed colocalization of fibroblasts with the inflammation pattern, suggesting a possible mechanism of fibroblast-induced inflammatory signaling, which was confirmed in a patient-derived co-culture organoid model.

With the high-dimensional insights derived from each layer of their experiment, Dr. Fertig’s team was able to provide a mechanistic hypothesis for their observations: fibroblasts seemed to induce inflammatory signaling within the precancerous epithelial compartment, causing a cell state transition that gave rise to the proliferative signature. Their approach represents an evolution in how researchers can begin to predict how a cancer is going to progress, or respond to therapy, based on the cellular and molecular composition of the precancerous lesion. This approach holds even greater promise in the clinical trial setting, where high-dimensional genomics data can be correlated to real patient outcomes.

Why Xenium In Situ for cancer research

As in situ approaches increase in prominence, it’s important to understand what kinds of questions cancer researchers can ask with high-dimensional, spatial transcriptomics profiling at single cell resolution. 10x’er Abbey Cutchin shared a few of the use cases scientists she spoke with at AACR 2023 were most interested in:

“[Scientists] are interested in looking at cellular interactions occurring at the sites of immune activation in cancer. In the case of inflammation, cells typically accumulate on top of each other quite densely, so you really need single cell resolution to resolve those cellular interactions. There’s interest in finding immune activation programs that are shared across different diseases, and describing what is happening in single cells spatially. The site of inflammation occurs in the tissue, so to be able to look at the spatial programming of these immune modules would be really valuable. I think many people would love to look at responders and non-responders to different immunotherapies, too, in order to resolve spatial biomarkers in the context of immunotherapy.”

It’s not just scientists that see the extreme significance of maximizing insights from precious patient samples with high-dimensional spatial analysis. Cutchin explained that one research team she spoke with at AACR was able to purchase a Xenium In Situ Analyzer through patient-donated funds: “That was really cool to hear—this patient, who has a devastating disease and is about to pass, donated funds for the Xenium. The research group will be studying glioblastoma, metastatic lesions in the colon, as well as lung and rare cancers.”

Keep exploring what cancer researchers can learn from precious tumor samples with Xenium In Situ analysis in this poster from AACR 2023, where we show data from an experiment to spatially map dozens of expressed mutations at subcellular resolution in FFPE breast cancer tissue sections.

Single cell and spatial multiomics resolve diverse cell states in archival FFPE and fresh frozen breast cancer tissue

In his presentation at AACR 2023, Alexander Swarbrick, PhD, Senior Principal Research Fellow and co-leader of the Strategic Program in Dynamic Cancer Ecosystems at the Garvan Institute of Medical Research, reinforced a foundational principle of cancer biology: cancers are complex tissues composed of diverse cell types and states whose multiplexed interactions within the tissue ecosystem define the clinical behavior of the tumor. In order to effectively intercept solid tumor progression, it is therefore essential to develop a detailed understanding of the cellular composition of the tumor and map cellular communities and interactions in a spatially resolved manner.

Dr. Swarbrick’s team took this approach to develop multimodal single cell and spatial atlases of breast cancer. In particular, they took advantage of new technologies to access archived clinical FFPE breast cancer samples, opening the door to many new experiments and questions. Describing a method his team developed called snPATHO-seq, which enables single nucleus RNA sequencing and Visium Spatial analysis from sequential sections of paraffin-embedded tissue, Dr. Swarbrick said:

“I think these technologies are going to be really important in oncology because it lets us do retrospective analysis. This might be, for example, going back and selecting patients with different outcomes from a clinical trial; or perhaps comparing a metastatic lesion to its primary lesion that was collected 5, 10 years earlier. This is now possible.”

Speaking to the potential of this method, which is enabled by Chromium Single Cell Gene Expression Flex, for even the most degraded samples, Dr. Swarbrick said, “A really tough sample—a metastatic lesion collected at autopsy from an individual—the RIN values are probably 1, but we can now generate really detailed single nuclei data representing many of the populations we would expect to be present in this lesion.” Dr. Swarbrick noted the importance of getting the most complete single cell data from degraded samples, as the single nuclei reference atlases they develop are then used to deconvolute spatial data and map cell types in tissue.

To capture the most diverse representation of cells in the complex breast cancer ecosystem, Dr. Swarbrick and his team have also turned to multiomic technologies that measure both the transcriptome and proteome in spatial context. They developed a method called SPITE-seq, which is built upon the Visium fresh frozen assay, to simultaneously map cellular transcripts and 50–100 proteins conjugated to barcoded antibodies in cancer tissue sections. This combined data can support sharper cellular clustering and resolve more unique cell states or phenotypes compared to results from the individual modalities alone. For example, they’ve used this approach to define a population of PD-L1Hi neutrophils in breast cancer tissue, suggesting neutrophils may be a sink for PD-L1 ligand at the center of tumors.

Dr. Swarbrick and his team have also been using these approaches to evaluate cancer cell heterogeneity, in particular mapping cancer-cell-specific gene expression modules in space to define recurring gene expression patterns between cells and cancer types, and to understand the relationships between these modular patterns in the tissue context. They noted that epithelial–mesenchymal transition (EMT) and proliferation patterns were regionally segregated in breast cancer tissue sections, hypothesizing that these regions could represent distinct genomic clones that give rise to different cancer lineages. To further test this hypothesis, the team used Visium CytAssist to map the transcriptome of large 11 x 11 mm archival FFPE breast cancer tissue sections that contained two malignant forms of breast cancer, ductal carcinoma in situ (DCIS) and infiltrating ductal carcinoma (IDC), and subsequently infer copy number variation from spatial data (InferCNV). This allowed them to define CNV gain or loss events in a regionally restricted manner, then associate those events with regional gene expression to ultimately understand the clonal relationships between divergent cancer populations in the same breast cancer sample.

The potential of cross-platform FFPE compatibility

Dr. Swarbrick’s experiments represent another key evolution of cancer research. As innovative single cell and spatial multiomics approaches unlock a greater depth of information in each sample, FFPE-compatible protocols also provide access to precious archival samples that have never been studied at single cell or spatial resolution before.

The larger cancer community at AACR 2023 was excited about the new questions and findings this expanded sample access could enable. Abbey Cutchin said, “A lot of the people we met at AACR were translational researchers at medical centers. This idea that now they could unlock large retrospective archival samples that they've had, that already have clinical outcome data, and really accelerate the pace at which they could define cohorts and ask questions—there was just a lot of excitement around this compatibility with FFPE.”

FFPE compatibility also holds promise to increase the scale of experiments aimed at screening precious clinical samples. We shared data from a multi-patient breast cancer tissue microarray analyzed with Visium at AACR 2023 that you can explore in this poster.

Technology unlocks a new way to study cancer 

“Our goal at 10x Genomics is to put a toolkit in the hands of researchers that enables them to ask questions at a scale and ambition that stand up to the scope of the challenge at hand—to match the complexity of cancer, the complexity and heterogeneity within individuals and between individuals, at different stages of disease. It's an enormous level of complexity,” Cutchin said.

With all of the amazing work showcased at AACR 2023, one thing is clear: innovative applications of high-dimensional single cell and spatial multiomics methods hold promise to fundamentally advance our understanding of cancer. Whether solid or hematological tumors, precancerous or malignant, treated or untreated, there’s something to learn from every precious sample—and we stand with the community of dedicated researchers that will leave no stone unturned, no question unasked, in their journey to improve therapies and outcomes for patients.

To learn more about the technologies discussed in this article, please explore our resources for:

Find more examples of how 10x Genomics tools apply to cancer research in our recent blogs. Learn how Xenium In Situ identified a small triple-positive region in a breast cancer sample that other methods missed; review the latest advancements in CAR T-cell therapy powered by single cell assays; and explore the value of spatial transcriptomics for cancer drug discovery.