Single-cell RNA sequencing (scRNA-seq) enables the systematic identification of cell populations in a tissue, but characterizing their spatial organization remains challenging. We combine a microarray-based spatial transcriptomics method that reveals spatial patterns of gene expression using an array of spots, each capturing the transcriptomes of multiple adjacent cells, with scRNA-Seq generated from the same sample. To annotate the precise cellular composition of distinct tissue regions, we introduce a method for multimodal intersection analysis. Applying multimodal intersection analysis to primary pancreatic tumors, we find that subpopulations of ductal cells, macrophages, dendritic cells and cancer cells have spatially restricted enrichments, as well as distinct coenrichments with other cell types. Furthermore, we identify colocalization of inflammatory fibroblasts and cancer cells expressing a stress-response gene module. Our approach for mapping the architecture of scRNA-seq-defined subpopulations can be applied to reveal the interactions inherent to complex tissues.
Subscribe to Journal
Get full journal access for 1 year
only $20.83 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
The scRNA-seq and ST data reported in this manuscript have been deposited to the Gene Expression Omnibus under accession number GSE111672.
Offit, K. A decade of discovery in cancer genomics. Nat. Rev. Clin. Oncol. 11, 632–634 (2014).
Kolodziejczyk, A. A., Kim, J. K., Svensson, V., Marioni, J. C. & Teichmann, S. A. The technology and biology of single-cell RNA sequencing. Mol. Cell 58, 610–620 (2015).
Papalexi, E. & Satija, R. Single-cell RNA sequencing to explore immune cell heterogeneity. Nat. Rev. Immunol. 18, 35–45 (2017).
Patel, A. P. et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science 344, 1396–1401 (2014).
Tirosh, I. et al. Single-cell RNA-seq supports a developmental hierarchy in human oligodendroglioma. Nature 539, 309–313 (2016).
Tirosh, I. et al. Dissecting the multicellular ecosystem of metastatic melanoma by single-cell RNA-seq. Science 352, 189–196 (2016).
Venteicher, A. S. et al. Decoupling genetics, lineages, and microenvironment in IDH-mutant gliomas by single-cell RNA-seq. Science (80-.) 355, eaai8478 (2017).
Darmanis, S. et al. Single-cell RNA-seq analysis of infiltrating neoplastic cells at the migrating front of human glioblastoma. Cell Rep. 21, 1399–1410 (2017).
Chung, W. et al. Single-cell RNA-seq enables comprehensive tumour and immune cell profiling in primary breast cancer. Nat. Commun. 8, 15081 (2017).
Horning, A. M. et al. Single-cell RNA-seq reveals a subpopulation of prostate cancer cells with enhanced cell cycle-related transcription and attenuated androgen response. Cancer Res. 78, 853–864 (2017).
Dirkse, A. et al. Stem cell-associated heterogeneity in glioblastoma results from intrinsic tumor plasticity shaped by the microenvironment. Nat. Commun. 10, 1787 (2019).
Filbin, M. G. et al. Developmental and oncogenic programs in H3K27M gliomas dissected by single-cell RNA-seq. Science (80-.) 360, 331–335 (2018).
Lawson, D. A. et al. Single-cell analysis reveals a stem-cell program in human metastatic breast cancer cells. Nature 526, 131–135 (2015).
Savage, P. et al. A targetable EGFR-dependent tumor-initiating program in breast cancer. Cell Rep. 21, 1140–1149 (2017).
Sharma, A. et al. Longitudinal single-cell RNA sequencing of patient-derived primary cells reveals drug-induced infidelity in stem cell hierarchy. Nat. Commun. 9, 4931 (2018).
Dalerba, P. et al. Single-cell dissection of transcriptional heterogeneity in human colon tumors. Nat. Biotechnol. 29, 1120–1127 (2011).
Puram, S. V. et al. Single-cell transcriptomic analysis of primary and metastatic tumor ecosystems in head and neck cancer. Cell 171, 1611–1624.e24 (2017).
Rambow, F. et al. Toward minimal residual disease-directed therapy in melanoma. Cell 174, 843–855 (2018).
Baron, M. et al. Cancer archetypes co-opt and adapt the transcriptional programs of existing cellular states. Preprint at bioRxiv https://doi.org/10.1101/396622 (2018).
Chen, K. H., Boettiger, A. N., Moffitt, J. R., Wang, S. & Zhuang, X. Spatially resolved, highly multiplexed RNA profiling in single cells. Science 348, aaa6090 (2015).
Shah, S., Lubeck, E., Zhou, W. & Cai, L. In situ transcription profiling of single cells reveals spatial organization of cells in the mouse hippocampus. Neuron 92, 342–357 (2016).
Long, X., Colonell, J., Wong, A. M., Singer, R. H. & Lionnet, T. Quantitative mRNA imaging throughout the entire drosophila brain. Nat. Methods 14, 703–706 (2017).
Lee, J. H. et al. Highly multiplexed subcellular RNA sequencing in situ. Science (80-.) 343, 1360–1363 (2014).
Nichterwitz, S. et al. Laser capture microscopy coupled with Smart-seq2 for precise spatial transcriptomic profiling. Nat. Commun. 7, 12139 (2016).
Achim, K. et al. High-throughput spatial mapping of single-cell RNA-seq data to tissue of origin. Nat. Biotechnol. 33, 503–509 (2015).
Satija, R., Farrell, J. A., Gennert, D., Schier, A. F. & Regev, A. Spatial reconstruction of single-cell gene expression data. Nat. Biotechnol. 33, 495–502 (2015).
Ståhl, P. L. et al. Visualization and analysis of gene expression in tissue sections by spatial transcriptomics. Science 353, 78–82 (2016).
Thrane, K., Eriksson, H., Maaskola, J., Hansson, J. & Lundeberg, J. Spatially resolved transcriptomics enables dissection of genetic heterogeneity in stage III cutaneous malignant melanoma. Cancer Res. 78, 5970–5979 (2018).
Berglund, E. et al. Spatial maps of prostate cancer transcriptomes reveal an unexplored landscape of heterogeneity. Nat. Commun. 9, 2419 (2018).
Lundmark, A. et al. Gene expression profiling of periodontitis-affected gingival tissue by spatial transcriptomics. Sci. Rep. 8, 9370 (2018).
Asp, M. et al. Spatial detection of fetal marker genes expressed at low level in adult human heart tissue. Sci. Rep. 7, 12941 (2017).
Maniatis, S. et al. Spatiotemporal dynamics of molecular pathology in amyotrophic lateral sclerosis. Science 364, 89–93 (2019).
Giacomello, S. et al. Spatially resolved transcriptome profiling in model plant species. Nat. Plants 3, 17061 (2017).
Junker, J. P. et al. Genome-wide RNA tomography in the zebrafish embryo. Cell 159, 662–675 (2014).
Chen, J. et al. Spatial transcriptomic analysis of cryosectioned tissue samples with Geo-seq. Nat. Protoc. 12, 566–580 (2017).
Saiselet, M. et al. Transcriptional output, cell types densities and normalization in spatial transcriptomics. Preprint at bioRxiv https://doi.org/10.1101/503870 (2018).
Wagner, F., Yan, Y. & Yanai, I. K-nearest neighbor smoothing for high-throughput single-cell RNA-Seq data. Preprint at bioRxiv https://doi.org/10.1101/217737 (2018).
Ziegenhain, C. et al. Comparative analysis of single-cell RNA sequencing methods. Mol. Cell 65, 631–643.e4 (2017).
Griffin, C. A. et al. Consistent chromosome abnormalities in adenocarcinoma of the pancreas. Cancer Res. 55, 2394–2399 (1995).
Shiga, K. et al. Cancer-associated fibroblasts: their characteristics and their roles in tumor growth. Cancers (Basel) 7, 2443–2458 (2015).
Motta, P. M., Macchiarelli, G., Nottola, S. A. & Correr, S. Histology of the exocrine pancreas. Microsc. Res. Tech. 37, 384–398 (1997).
May, D. et al. Ero1-Lα plays a key role in a HIF-1-mediated pathway to improve disulfide bond formation and VEGF secretion under hypoxia: implication for cancer. Oncogene 24, 1011–1020 (2005).
Sedlakova, O. et al. Carbonic anhydrase IX, a hypoxia-induced catalytic component of the pH regulating machinery in tumors. Front. Physiol. 4, 400 (2014).
Baron, M. et al. A single-cell transcriptomic map of the human and mouse pancreas reveals inter- and intra-cell population structure. Cell Syst. 3, 346–360 (2016).
Mehrfeld, C., Zenner, S., Kornek, M. & Lukacs-Kornek, V. The contribution of non-professional antigen-presenting cells to immunity and tolerance in the liver. Front. Immunol. 9, 635 (2018).
Wosen, J. E., Mukhopadhyay, D., Macaubas, C. & Mellins, E. D. Epithelial MHC class II expression and its role in antigen presentation in the gastrointestinal and respiratory tracts. Front. Immunol. 9, 2144 (2018).
Rőszer, T. Understanding the mysterious M2 macrophage through activation markers and effector mechanisms. Mediators Inflamm. 2015, 816460 (2015).
Sanyal, R. et al. MS4A4A: a novel cell surface marker for M2 macrophages and plasma cells. Immunol. Cell Biol. 95, 611–619 (2017).
Lopez-Bergami, P., Lau, E. & Ronai, Z. Emerging roles of ATF2 and the dynamic AP1 network in cancer. Nat. Rev. Cancer 10, 65–76 (2010).
Hyakusoku, H. et al. JunB promotes cell invasion, migration and distant metastasis of head and neck squamous cell carcinoma. J. Exp. Clin. Cancer Res. 35, 6 (2016).
Öhlund, D. et al. Distinct populations of inflammatory fibroblasts and myofibroblasts in pancreatic cancer. J. Exp. Med. 214, 579–596 (2017).
Obeid, J. M., Hu, Y., Erdag, G., Leick, K. M. & Slingluff, C. L. The heterogeneity of tumor-infiltrating CD8+ T cells in metastatic melanoma distorts their quantification. Melanoma Res. 27, 211–217 (2017).
Welch, J. D. et al. Single-cell multi-omic integration compares and contrasts features of brain cell identity. Cell 177, 1873–1887.e17 (2019).
Stuart, T. et al. Comprehensive integration of single-cell data. Cell 177, 1888–1902.e21 (2019).
Ligorio, M. et al. Stromal microenvironment shapes the intratumoral architecture of pancreatic cancer. Cell 178, 160–175.e27 (2019).
Elyada, E. et al. Cross-species single-cell analysis of pancreatic ductal adenocarcinoma reveals antigen-presenting cancer-associated fibroblasts. Cancer Discov. 9, 1102–1123 (2019).
Akira, S. et al. A nuclear factor for IL-6 expression (NF-IL6) is a member of a C/EBP family. EMBO J. 9, 1897–1906 (1990).
Schuringa, J.-J., Timmer, H., Luttickhuizen, D., Vellenga, E. & Kruijer, W. c-Jun and c-Fos cooperate with Stat3 in IL-6-induced transactivation of the IL-6 response element (IRE). Cytokine 14, 78–87 (2001).
Venugopal, R. & Jaiswal, A. K. Nrf2 and Nrf1 in association with Jun proteins regulate antioxidant response element-mediated expression and coordinated induction of genes encoding detoxifying enzymes. Oncogene 17, 3145–3156 (1998).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Van Der Maaten, L. & Hinton, G. Visualizing data using t-SNE. J. Mach. Learn. Res. 9, 2579–2605 (2008).
Hashimshony, T. et al. CEL-Seq2: sensitive highly-multiplexed single-cell RNA-Seq. Genome. Biol. 17, 77 (2016).
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Anders, S., Pyl, P. T. & Huber, W. HTSeq—a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015).
Gaujoux, R. & Seoighe, C. A flexible R package for nonnegative matrix factorization. BMC Bioinformatics 11, 367 (2010).
Carmona-Saez, P., Pascual-Marqui, R. D., Tirado, F., Carazo, J. M. & Pascual-Montano, A. Biclustering of gene expression data by non-smooth non-negative matrix factorization. BMC Bioinformatics 7, 78 (2006).
We thank C. Loomis, Z. Dewan and B. Dabovic from the NYU Experimental Pathology Core, and B. Zeck and L. Chiriboga from the NYU Center for Biospecimen Research and Development for technical assistance, A. Weil from the NYU CBRD for sample acquisition, and members of the Yanai laboratory for constructive comments.
I.Y. is a shareholder of OneCell Medical Ltd.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Integrated supplementary information
(a-b) Histogram of unique transcripts detected per cell (log10) in PDAC-A (a) and PDAC-B (b). (c-d) Histogram of unique genes expressed per cell (log10) in PDAC-A (c) and PDAC-B. (e-f) Histogram of unique transcripts expressed per gene across all cells in PDAC-A (e) and PDAC-B (f). (g-h) PDAC-A (g) and PDAC-B (h) cells colored by UMIs per cell (log10) in tSNE space.
(a) PDAC-A (top) and PDAC-B (bottom) CNV profiles inferred from scRNA-Seq (same as Fig. 1d, e). A subset of 200 randomly selected ductal cells were chosen as a negative control for the analysis and analyzed together with cancer clusters. (b) Expression levels of KRT19 (marker of malignant and non-malignant ductal epithelial cells), TM4SF1, and S100A4 projected onto t-SNE of PDAC-A (top) and PDAC-B (bottom). Note specificity of TM4SF1 expression for PDAC-A cancer cluster 1, and S100A4 expression for PDAC-A cancer cluster 2. In PDAC-B, TM4SF1 is expressed primarily by cancer cells whereas an S100A4 expressing cancer population is absent. (c) Double immunofluorescence staining (n = 2) for KRT19 and TM4SF1 in PDAC-A FFPE tissue. Note co-localization of KRT19 and TM4SF1 signal in malignant ducts (top panel, white arrowheads), but not in non-malignant ducts (bottom panel, white arrowheads). Scale bar, 250 µm. (d) Double immunofluorescence staining (n = 2) for KRT19 and S100A4 in PDAC-A FFPE tissue. Note co-localization of KRT19 and S100A4 signal in malignant ducts (top panel, white arrowheads), but not in non-malignant ducts (bottom panel, white arrowheads). Scale bar, 250 µm.
(a) Overview of PDAC-A tissue histology. Scale bar, 1mm. (b) Inset of pancreatic cancer ducts and surrounding desmoplasia. White arrowheads indicate cancer cells. Black arrowheads show the surrounding stroma and desmoplasia. Scale bar, 100 µm. (c) Inset of pancreatic tissue. Arrowheads indicate the acini. Scale bar, 100 µm. (d) Inset of duct epithelium and inflamed tissue. White arrowheads indicate the pancreatic ducts and the black arrowheads point to inflammatory cells with smaller nuclei. Scale bar, 100 µm. (e) Overview of PDAC-B tissue histology. Scale bar, 1mm. 3 ST replicates were generated from this tumor. (f) Inset of PDAC-B tissue showing normal ducts (white arrow) surrounded by interstitial space (black arrows). Scale bar, 100 µm.
ST spots were mapped back onto the H&E stained tissue, and brightfield images were extracted from the location of each ST spot. In each enlarged spot, the dark purple dots are nuclei, and the purple background is the cytoplasm and extracellular space. Enlarged spots shown demonstrate that ST spots can capture as few as ten or less cells and as many as a few dozen cells. Scale bar, 1 mm.
(a-b) Histogram of unique transcripts detected per spot (log10) in PDAC-A (a) and PDAC-B (b). (c-d) Histogram of unique genes detected per spot (log10) in PDAC-A (c) and PDAC-B (d). (e-f) Distribution of unique transcripts detected across ST spots in PDAC-A (e) and PDAC-B (f). (g-h) PC scores for the first six components projected onto ST data for PDAC-A (g) and PDAC-B (h). (i) Heatmap of MIA enrichment values with the indicated range of p-value thresholds for selecting cell type (scRNA-Seq) and tissue region (ST) specific genes after a two-sided t-test for marker gene selection (see Methods). Asterisk indicates the enrichment of fibroblast-specific genes in the PDAC-A cancer region with the corresponding p-value thresholds used in Fig. 2h. The values plotted in the heatmap were calculated using the hypergeometric distribution with 19,738 genes used as the background. (j) Same as (i), for varied p-value thresholds for cancer cluster 1 markers and PDAC-A pancreatic tissue region markers. Asterisk indicates the depletion of cancer cluster 1 specific genes in the PDAC-A pancreatic tissue region with the corresponding p-value thresholds used in Fig. 2h. (k) Boxplots of hypergeometric enrichment (P-value) of the fibroblast-specific genes with different sized down-samplings of the ST cancer-region gene set. Downsampling of cancer region-specific genes (of the 1734 genes with P< 0.05) was for 750, 500, 250, 100, 50, 25, and 5 genes. Each box plot represents the enrichment p-value of the fibroblast-specific genes with 500 independent random subsets of the cancer region-specific genes. The red central mark for each box plot represents the median enrichment.
FFPE tissue was co-stained for KRT19 (duct marker) and subpopulation markers CA9 (a), AQP3 (b), HLA-DR (c), and TFF1 (d), as shown in Fig. 3. Individual signals are shown here in addition to the merged signal to better demonstrate marker co-localization (n = 2). Scale bar, 250 µm.
(a) Expression of M1 (top row) and M2 (bottom row) marker genes (two-tailed Student’s t-test, P < 10−5). Center of violin plots show the median expression level, and grey lines show interquartile range. Violin plots for M1 and M2 macrophages represent 19 and 21 data points, respectively. (b) Enrichment of macrophage subpopulations across PDAC-A ST regions. Indicated are the number of genes used for MIA. Values plotted are calculated using the hypergeometric distribution with 19,738 genes in the background. The number of subpopulation and region-specific genes used in the calculation are shown. (c) Expression of dendritic cells A (top panel) or B (bottom panel) marker genes (two-tailed Student’s t-test, P < 10−5). Center of violin plots show the median expression level, and grey lines show interquartile range. Violin plots for A and B dendritic cell subpopulations represent 12 and 33 data points, respectively. (d) Enrichment of dendritic cell subpopulations across PDAC-A ST regions. Indicated are the number of genes used for MIA. Values plotted are calculated using the hypergeometric distribution with 19,738 genes in the background. The number of subpopulation and region-specific genes used in the calculation are shown.
(a) Tissue histology for PDAC-A-2 and PDAC-A-3 used for ST. Highlighted in red are the cancer cells in the tissue. Scale bar, 1 mm. 3 ST replicates were generated from this PDAC-A. (b) Distribution of UMIs across the PDAC-A-2 and PDAC-A-3 ST tissue. (c) ST data statistics for PDAC-A-2 and PDAC-A-3 samples. Top, histogram of UMIs per spot. Bottom, histogram of genes expressed per spot. (d) MIA map of the PDAC-A-2 tissue regions. Color bar above MIA map indicates the clustering assignments. The hypergeometric distribution was used to calculate p-values shown with 19,738 genes in the background. (e) Hypergeometric between gene sets specific to PDAC-A-1 and PDAC-A-2 tissue regions. Color bar on left corresponds to PDAC-A-1 ST clustering assignments, and color bar on top corresponds to PDAC-A-2 ST clustering assignments.
(a) Histogram of unique transcripts (left), genes expressed (middle), and unique transcripts per gene per cell in PDAC-C. (b) t-SNE visualization of PDAC-C cell type annotations by hierarchical clustering. (c) Standardized expression of KRT19 (left) and TM4SF1 (right) projected onto t-SNE. (d) CNV profiles inferred from scRNA-Seq. A subset of 200 randomly selected ductal cells were chosen as a negative control for the analysis and analyzed together with cancer clusters. All remaining cells in the dataset were used as the reference for CNV inference.
(a) H&E staining of two additional ST samples from PDAC-B tumor, and four additional ST samples each originating from different patients. Brightfield images were annotated for the presence of cancer cells (outlined in red). Scale bar, 1 mm. (b) Histograms of unique transcripts detected per spot (log10) for the indicated ST datasets. (c) Histograms of genes expressed per spot (log10) for the indicated ST datasets.
(a-f) ST spots were analyzed using PCA followed by hierarchical clustering of PC scores, and genes specific to each region were identified for MIA (see Methods). MIA was performed using scRNA-Seq defined marker genes to determine which ST cluster represented the cancer region in each respective ST dataset. Color bars above each MIA map represent ST clustering assignments. Asterisks indicate the ST cancer cluster used for analysis in Fig. 5c. The hypergeometric distribution was used to calculate p-values using the number of region specific genes shown and with 19,738 background genes.
(a) t-SNE visualization of metastatic melanoma scRNA-Seq data from Tirosh et al 2016b6. Plots are colored by patient (left panel) and by clustering assignments (right panel). (b) Marker genes of T-cell subsets. Shown are expression levels of marker genes for CD8+ T-cells (top row), CD4+ regulatory T-cells (middle row) and helper T-cells (bottom row). Center of violin plots show the median expression level, and grey lines show interquartile range. Violin plots for CD8+, CD4+ T-regs, and T-helper cells represent 553, 193, and 368 data points, respectively. (c-d) Metastatic melanoma ST data from two patients published in Thrane et al 201828. Top left panels show histological annotations of H&E stained tissue, and top right panels show clustering assignments of the ST data. For the ST data shown, the ST spots are colored according to their clustering assignments. Bottom panels show MIA maps of the respective metastatic melanoma ST datasets. Marker genes were identified using the scRNA-Seq data from Tirosh et al 2016b6, and MIA was applied to generate cell population – tissue regions maps of ST data from Thrane et al 201828.
Supplementary Figs. 1–12.
Average number of transcripts per ST spot reported in other publications. The table indicates the number of transcripts detected per spot in other studies using the ST method. Included are only publications that report the average number of detected transcripts
About this article
Cite this article
Moncada, R., Barkley, D., Wagner, F. et al. Integrating microarray-based spatial transcriptomics and single-cell RNA-seq reveals tissue architecture in pancreatic ductal adenocarcinomas. Nat Biotechnol (2020). https://doi.org/10.1038/s41587-019-0392-8