Some of our favourite Visium spatial transcriptomics publications from 2023

From identifying a previously undescribed T-cell state in cancer, to defining the genes involved in healthy palate development, our 2023 publication picks cover the most innovative and interesting applications of the Visium Spatial platform. And they capture a recurring theme: spatial discovery powered by unbiased insights. 

With the ability to spatially resolve the whole transcriptome across a complete tissue section, Visium gives researchers an unbiased view of the biology of their samples. This can reveal insights that researchers weren’t looking for; generate hypotheses that can be further tested and validated; and multiply the discovery potential in every sample. 

Novel T-cell state confirmed across cancer types

In their May 2023 Nature Medicine publication (1), researchers from the University of Texas MD Anderson Cancer Center established the most detailed picture of heterogeneous T cells in the tumor microenvironment (TME) to date. Their cell atlas included 308,048 single T-cell transcriptomes from 16 cancer types, taken from 375 human patient samples. This revealed 32 distinct T-cell states, including a previously undescribed T-cell stress response state (TSTR) associated with increased expression of heat shock genes HSPA1A/HSPA1B. 

To explore this T-cell state and its spatial relationships in tissue, the team used Visium spatial transcriptomics on 33 FFPE tissue sections from six cancer types: melanoma, lung adenocarcinoma, breast cancer, cutaneous squamous cell carcinoma, clear cell renal cell carcinoma, and gastric adenocarcinoma. They found TSTR cells mainly within lymphocyte aggregates, potential tertiary lymphoid structures in tumor beds, or around tumor edges. Clinical data showed TSTR cells increased after immunotherapy and were more prevalent in non-responders, suggesting they are less effective at fighting cancer. These findings suggest that TSTR cells, located in lymphocyte-rich regions of the tumor bed, may interact with other immune cells and influence immunotherapy resistance.

Unreported genes discovered in the fusing palate transcriptome

According to a 2019 study, 1 in every 1,700 babies is born with cleft palate in the United States (2). Cleft palate is a developmental defect that occurs when the tissue that makes up the roof of the mouth does not join together completely during pregnancy (2), making it all the more important to understand the cellular and molecular mechanisms underlying healthy palate development.  

In their Nature Communications publication (3), a team from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) at the NIH used Visium analysis on FFPE coronal mid-palatal sections of embryonic mice. By examining stages where palate shelves contact (E14.5) and fuse (E15.5), they mapped the active genes driving normal palate fusion and identified differentially expressed genes between these stages.

“With an eye toward discovery, we turned our attention to the most highly differentially expressed genes identified from the spRNA-seq analysis to identify potential markers of palatal fusion across space and time.” —Piña et al. (3)

Figure from Piña J, et al. (3). Mid-palatal coronal cross sections of whole embryo heads were placed on barcoded Visium slides. A) In vivo clusters were defined from the whole embryo head, demonstrating spatial relationships and morphogenetic diversity of expression, further filtered for only those clusters encoded on the barcodes placed within the palate tissue in each respective section to identify top differentially expressed genes (DEGs) B) from E14.5 vs. E15.5 in the palate. C) Spatial gene expression feature plots for the three enriched genes identified. D) Colored circles correspond to 55 μm-diameter Visium transcriptomic resolution. Increased expression levels, delineated using the 10x Genomics Loupe Browser, are represented here with darker shades of green (Deup1), blue (Dynlrb2), or red (Lrrc23). The combined localization of these genes is indicated by overlapping concentric colored circles, the diameter of which does not correspond to degree of expression. nc nasal cavity, oc oral cavity, t tongue, mes midline epithelial seam; scale bar: 200 μm. CREDIT: Piña J, et al. (3). (CC BY 4.0)

Their unbiased, whole transcriptome Visium experiment uncovered three previously unreported genes involved in palate fusion—Deup1, Dynlrb2, and Lrrc23—all associated with cilia function, and localized in the palate ciliated nasal epithelium at E15.5. This suggests that nasal epithelium cells may play a crucial role in facilitating cell migration during the fusion of the palate shelves. Detailed insights like these into the temporal and spatial cellular dynamics of palate fusion represent a significant step toward, in the author’s words, “new preclinical models of palate ossification and submucosal clefts, paving the way toward potential therapies to correct cleft palate defects in humans.” 

Revealing transcriptomic and metabolic changes in the injured human brain

Brain injuries are scary and, in many cases, life-altering. Studies recorded through the Traumatic Brain Injury Model Systems National Database showed that, for a US population of 16 years and older that received inpatient rehabilitation services after moderate to severe traumatic brain injury (TBI), 26% saw positive improvement at the five-year mark after injury—however, 30% got worse, while 22% had died (4). 

Researchers from the Shanghai Pudong New Area People’s Hospital wanted to explore the underlying biological changes that might be associated with varying therapeutic outcomes for people with moderate to severe TBI. They were specifically looking to understand the relationship between transcriptomic and metabolic changes in injured brain tissue. Metabolic readouts—molecules such as lipids or proteins left behind from cellular processes—can provide a snapshot of the physiology of cells or an organ, and potentially reveal novel biomarkers or functional insights when paired with transcriptomic data. 

The team analyzed surgical brain samples from six people with moderate or severe TBIs using Visium Spatial Gene Expression and airflow-assisted desorption electrospray ionization mass spectrometry imaging (AFADESI-MSI)—a spatial metabolomics method. Combining transcriptomic and metabolic data revealed upregulated expression of metabolite-associated genes, SEPTIN4 and S1PR5, within the site of injury in severe TBI groups. Specifically, SEPTIN4, which is typically involved in cell death, localized to regions with neuronal loss; and S1PR5, which is involved in lipid metabolism processes, localized to tissue regions with high abundance of a lipid marker (5). 

These findings suggest increased lipid peroxidation impairs synapse formation and neuron development. The data also provides a distinguishing molecular pattern between moderate and severe TBI, and adds to our understanding of why patients with severe TBI tend to have poorer neurological outcomes and responses to comprehensive therapies. 

Revealing the drivers of CD4⁺ T-cell lineage commitment

While we know that CD4⁺ T cells are important—they differentiate into diverse T-cell subtypes that play a critical role in activating and regulating immune responses to pathogens (6,7)—researchers from the NIH and University of Pennsylvania Medical School wanted to fill the gaps in our understanding of the factors that control commitment to the CD4⁺ lineage in the thymus. 

In their Science Immunology publication (6), the team took a multidimensional approach to studying the transcription factors (TFs) that drive the split from double-positive CD4⁺CD8⁺ T cells to the CD4⁺ lineage during T-cell selection. Prior studies suggested that multiple factors play a part in initiating the gene expression program characteristic of the CD4⁺ lineage, directing the team to establish a double knockout (KO) mouse model of two relatively unexplored genes, Zfp281 and Zfp148. These were putative activators of Thpok expression, a gene encoding a TF that in turn prevents expression of, and chromatin accessibility at, CD8⁺ lineage genes, thus guiding the course of differentiation towards the CD4⁺ lineage. 

Initially using single-cell RNA-seq, researchers demonstrated that knocking out Zfp281 and Zfp148 altered transcriptomic profiles in immature CD4+ thymocytes, resulting in lower CD4+ lineage scores compared to controls. Subsequent spatial analysis with Visium CytAssist Spatial Gene Expression on FFPE tissue sections confirmed reduced Ccr7 gene expression in KO samples, affecting T-cell migration within the thymus. This spatial analysis also revealed a significant decrease in CD4+ lineage signatures in the KO medulla compared to controls, highlighting the initial impact of Zfp281 and Zfp148 on CD4+ T-cell differentiation within thymus tissue.

Histology of the thymus. The structure of thymus in young humans shows differentiation into lightly staining medulla and more densely packed darker cortex. The thymus is divided into larger lobules with septa consisting of connective tissue and wrapped in a connective tissue capsule. Originally published by Hill et al. (Hill 2019). Credit: Klocperk A. Adaptive immune system in patients with primary immunodeficiencies. (2019). doi: 10.13140/RG.2.2.22888.42246.

Searching for transcript isoforms in spatial context

The discovery power of whole transcriptome spatial gene expression can be strengthened with technological innovations that allow researchers to obtain a more complete picture of biological complexity in a tissue section. That includes transcript isoforms—sequences of RNA transcribed from the same gene but combined in various ways to produce different, but related, mRNA transcripts—that may be hiding in plain sight.  

Researchers from the Institut de Pharmacologie Moléculaire et Cellulaire and the KTH Royal Institute of Technology validated a method called spatial isoform transcriptomics (SiT) in their Nucleic Acids Research publication (8). The technique builds on the Visium assay, employing Nanopore long-read sequencing to build spatial gene expression libraries that capture full-length mRNA sequences to detect transcript isoforms. Using this approach, they studied different regions of mouse brain, including the olfactory bulb, to explore transcript isoform diversity. 

Within the olfactory bulb, they observed regional isoform switching for 36 genes, one of the most prominent of which was Plp1 (proteolipid protein 1). Implicated in central nervous system pathologies such as dysmyelination, Plp1 isoforms were differentially expressed between the inner granule cell layer (where the full isoform, PLP, was preferentially expressed) and the outer regions of the olfactory nerve layer (where the shortened DM20 isoform was preferentially expressed). 

The SiT method allowed the researchers to identify 35 additional codons present in PLP and quantify the balance of both isoforms with spatial context. To further define the cellular source of Plp1 isoform expression, the team used an existing olfactory bulb single cell RNA-seq dataset to deconvolute the cell types in spatial spots defined by SiT. Their data suggested that myelinating-oligodendrocyte cells in the granule cell layer were the main source of the Plp1 standard isoform, while olfactory ensheathing cells in the olfactory nerve layer sourced the shorter DM20 isoform. Together, these findings offer new mechanistic insights into an isoform ratio implicated in a rare neurodegenerative disease, Pelizaeus-Merzbacher disease (8). 

Additional publications and blogs: 

• Benotmane J, et al. High-sensitive spatially resolved T cell receptor sequencing with SPTCR-seq. Nat Commun 14: 7432 (2023). doi: 10.1038/s41467-023-43201-6
• Researchers map the crucial cellular processes in early human pregnancy
• Researchers uncover how people with paralysis may walk again
• Novel lung immune niche provides long-term antibody hideouts
• Spatial and single cell tools open the black box of early human development
• Spatial transcriptomics—the next evolution of cancer drug discovery
• Simplifying spatial transcriptomics protocols with Visium CytAssist

References: 

1. Chu Y, et al. Pan-cancer T cell atlas links a cellular stress response state to immunotherapy resistance. Nat Med 29: 1550–1562 (2023). doi: 10.1038/s41591-023-02371-y
2. Facts about cleft lip and cleft palate. https://www.cdc.gov/ncbddd/birthdefects/cleftlip.html#ref
3. Piña J, et al. Multimodal spatiotemporal transcriptomic resolution of embryonic palate osteogenesis. Nat Commun 14: 5687 (2023). doi: 10.1038/s41467-023-41349-9
4. Moderate and severe TBI. https://www.cdc.gov/traumaticbraininjury/moderate-severe/index.html
5. Zheng P, et al. Integrated spatial transcriptome and metabolism study reveals metabolic heterogeneity in human injured brain. Cell Rep Med 4: 101057 (2023). doi: 10.1016/j.xcrm.2023.101057
6. Chopp L, et al. Zfp281 and Zfp148 control CD4+ T cell thymic development and TH2 functions. Sci Immunol 8: eadi9066 (2023). doi: 10.1126/sciimmunol.adi9066
7. Luckheeram R, et al. CD4+ T cells: Differentiation and functions. Clin Dev Immunol 2012: 925135 (2012). doi: 10.1155/2012/925135
8. Lebrigand K, et al. The spatial landscape of gene expression isoforms in tissue sections. Nucleic Acids Res 51:e47 (2023). doi: 10.1093/nar/gkad169

Source: Some of our favorite Visium spatial transcriptomics publications from 2023