Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Dissecting cellular crosstalk by sequencing physically interacting cells

Abstract

Crosstalk between neighboring cells underlies many biological processes, including cell signaling, proliferation and differentiation. Current single-cell genomic technologies profile each cell separately after tissue dissociation, losing information on cell–cell interactions. In the present study, we present an approach for sequencing physically interacting cells (PIC-seq), which combines cell sorting of physically interacting cells (PICs) with single-cell RNA-sequencing. Using computational modeling, PIC-seq systematically maps in situ cellular interactions and characterizes their molecular crosstalk. We apply PIC-seq to interrogate diverse interactions including immune–epithelial PICs in neonatal murine lungs. Focusing on interactions between T cells and dendritic cells (DCs) in vitro and in vivo, we map T cell–DC interaction preferences, and discover regulatory T cells as a major T cell subtype interacting with DCs in mouse draining lymph nodes. Analysis of T cell–DC pairs reveals an interaction-specific program between pathogen-presenting migratory DCs and T cells. PIC-seq provides a direct and broadly applicable technology to characterize intercellular interaction-specific pathways at high resolution.

This is a preview of subscription content, access via your institution

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Detection and sequencing of PICs.
Fig. 2: T cells interacting with PICs show early activation and differentiation.
Fig. 3: Studying T cell–myeloid interactions in an in vivo infection model.
Fig. 4: Regulatory T cells have a high capacity to interact with APCs.

Similar content being viewed by others

Data availability

RNA-seq data that support the findings of the present study were deposited in the Gene Expression Omnibus under accession code GSE135382. Source data for Figs. 3f and 4i and Supplementary Figs. 2 and 6 are presented with the paper. All other data supporting the findings of the present study are available from the corresponding author on reasonable request.

Code availability

The PIC-seq algorithm, including scripts and auxiliary data needed to reconstruct analysis files from count matrices to full figures, is provided as Supplementary Software.

References

  1. Banchereau, J. & Steinman, R. M. Dendritic cells and the control of immunity. Nature 392, 245–252 (1998).

    CAS  PubMed  Google Scholar 

  2. Freeman, S. A. & Grinstein, S. Phagocytosis: receptors, signal integration, and the cytoskeleton. Immunol. Rev. 262, 193–215 (2014).

    CAS  PubMed  Google Scholar 

  3. Lanier, L. L. Natural killer cell receptor signaling. Curr. Opin. Immunol. 15, 308–314 (2003).

    CAS  PubMed  Google Scholar 

  4. Jaitin, D. A. et al. Lipid-associated macrophages control metabolic homeostasis in a Trem2-dependent manner. Cell 178, 686–698 e614 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  5. Li, H. et al. Dysfunctional CD8 T cells form a proliferative, dynamically regulated compartment within human melanoma. Cell 176, 775–789 e718 (2019).

    CAS  PubMed  Google Scholar 

  6. Ledergor, G. et al. Single cell dissection of plasma cell heterogeneity in symptomatic and asymptomatic myeloma. Nat. Med. 24, 1867–1876 (2018).

    CAS  PubMed  Google Scholar 

  7. Gerber, T. et al. Single-cell analysis uncovers convergence of cell identities during axolotl limb regeneration. Science 362, eaaq0681 (2018).

  8. Keren-Shaul, H. et al. A unique microglia type associated with restricting development of Alzheimer’s disease. Cell 169, 1276–1290 e1217 (2017).

    CAS  PubMed  Google Scholar 

  9. Steinman, R. M. & Banchereau, J. Taking dendritic cells into medicine. Nature 449, 419–426 (2007).

    CAS  PubMed  Google Scholar 

  10. Zhu, J., Yamane, H. & Paul, W. E. Differentiation of effector CD4 T cell populations (*). Annu. Rev. Immunol. 28, 445–489 (2010). .

  11. Binnewies, M. et al. Unleashing type-2 dendritic cells to drive protective antitumor CD4+ T cell immunity. Cell 177, 556–571 e516 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  12. van Panhuys, N. Studying dendritic cell–T cell interactions under in vivo conditions. Methods Mol. Biol. 1584, 569–583 (2017).

    PubMed  Google Scholar 

  13. Barnden, M. J., Allison, J., Heath, W. R. & Carbone, F. R. Defective TCR expression in transgenic mice constructed using cDNA-based alpha- and beta-chain genes under the control of heterologous regulatory elements. Immunol. Cell Biol. 76, 34–40 (1998).

    CAS  PubMed  Google Scholar 

  14. Macosko, E. Z. et al. Highly parallel genome-wide expression profiling of individual cells using nanoliter droplets. Cell 161, 1202–1214 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  15. Klein, A. M. et al. Droplet barcoding for single-cell transcriptomics applied to embryonic stem cells. Cell 161, 1187–1201 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  16. Jaitin, D. A. et al. Massively parallel single-cell RNA-seq for marker-free decomposition of tissues into cell types. Science 343, 776–779 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  17. Rodriques, S. G. et al. Slide-seq: a scalable technology for measuring genome-wide expression at high spatial resolution. Science 363, 1463–1467 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  18. Eng, C. L. et al. Transcriptome-scale super-resolved imaging in tissues by RNA seqFISH. Nature 568, 235–239 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  19. Keren, L. et al. A structured tumor-immune microenvironment in triple negative breast cancer revealed by multiplexed ion beam imaging. Cell 174, 1373–1387 e1319 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  20. Wang, X. et al. Three-dimensional intact-tissue sequencing of single-cell transcriptional states. Science 361, eaat5691 (2018).

  21. Medaglia, C. et al. Spatial reconstruction of immune niches by combining photoactivatable reporters and scRNA-seq. Science 358, 1622–1626 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  22. Chen, K. H., Boettiger, A. N., Moffitt, J. R., Wang, S. & Zhuang, X. RNA imaging. Spatially resolved, highly multiplexed RNA profiling in single cells. Science 348, aaa6090 (2015).

    PubMed  PubMed Central  Google Scholar 

  23. Vento-Tormo, R. et al. Single-cell reconstruction of the early maternal-fetal interface in humans. Nature 563, 347–353 (2018).

    CAS  PubMed  Google Scholar 

  24. Cohen, M. et al. Lung single-cell signaling interaction map reveals basophil role in macrophage imprinting. Cell 175, 1031–1044 e1018 (2018).

    CAS  PubMed  Google Scholar 

  25. Camp, J. G. et al. Multilineage communication regulates human liver bud development from pluripotency. Nature 546, 533–538 (2017).

    CAS  PubMed  Google Scholar 

  26. Szczerba, B. M. et al. Neutrophils escort circulating tumour cells to enable cell cycle progression. Nature 566, 553–557 (2019).

    CAS  PubMed  Google Scholar 

  27. Halpern, K. B. et al. Paired-cell sequencing enables spatial gene expression mapping of liver endothelial cells. Nat. Biotechnol. 36, 962–970 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  28. Boisset, J. C. et al. Mapping the physical network of cellular interactions. Nat. Methods 15, 547–553 (2018).

    CAS  PubMed  Google Scholar 

  29. Keren-Shaul, H. et al. MARS-seq2.0: an experimental and analytical pipeline for indexed sorting combined with single-cell RNA sequencing. Nat. Protoc. 14, 1841–1862 (2019).

    CAS  PubMed  Google Scholar 

  30. Celli, S., Lemaitre, F. & Bousso, P. Real-time manipulation of T cell–dendritic cell interactions in vivo reveals the importance of prolonged contacts for CD4+ T cell activation. Immunity 27, 625–634 (2007).

    CAS  PubMed  Google Scholar 

  31. Kane, L. P., Lin, J. & Weiss, A. Signal transduction by the TCR for antigen. Curr. Opin. Immunol. 12, 242–249 (2000).

    CAS  PubMed  Google Scholar 

  32. Baran, Y. et al. MetaCell: analysis of single-cell RNA-seq data using K-nn graph partitions. Genome Biol. 20, 206 (2019).

    PubMed  PubMed Central  Google Scholar 

  33. Treutlein, B. et al. Reconstructing lineage hierarchies of the distal lung epithelium using single-cell RNA-seq. Nature 509, 371–375 (2014).

    CAS  PubMed  PubMed Central  Google Scholar 

  34. Bird, A. D. et al. Identification of glucocorticoid-regulated genes that control cell proliferation during murine respiratory development. J. Physiol. 585, 187–201 (2007).

    CAS  PubMed  PubMed Central  Google Scholar 

  35. Itani, O. A., Liu, K. Z., Cornish, K. L., Campbell, J. R. & Thomas, C. P. Glucocorticoids stimulate human sgk1 gene expression by activation of a GRE in its 5’-flanking region. Am. J. Physiol. Endocrinol. Metab. 283, E971–E979 (2002).

    CAS  PubMed  Google Scholar 

  36. Binder, E. B. The role of FKBP5, a co-chaperone of the glucocorticoid receptor in the pathogenesis and therapy of affective and anxiety disorders. Psychoneuroendocrinology 34(Suppl 1), S186–S195 (2009).

    CAS  PubMed  Google Scholar 

  37. Bonifaz, L. C. et al. In vivo targeting of antigens to maturing dendritic cells via the DEC-205 receptor improves T cell vaccination. J. Exp. Med. 199, 815–824 (2004).

    CAS  PubMed  PubMed Central  Google Scholar 

  38. Blecher-Gonen, R. et al. Single-cell analysis of diverse pathogen responses defines a molecular roadmap for generating antigen-specific immunity. Cell Syst. 8, 109–121 e106 (2019).

    CAS  PubMed  Google Scholar 

  39. Connor, L. M., Tang, S. C., Camberis, M., Le Gros, G. & Ronchese, F. Helminth-conditioned dendritic cells prime CD4+ T cells to IL-4 production in vivo. J. Immunol. 193, 2709–2717 (2014).

    CAS  PubMed  Google Scholar 

  40. Merad, M., Sathe, P., Helft, J., Miller, J. & Mortha, A. The dendritic cell lineage: ontogeny and function of dendritic cells and their subsets in the steady state and the inflamed setting. Annu. Rev. Immunol. 31, 563–604 (2013).

    CAS  PubMed  Google Scholar 

  41. Fujii, S., Liu, K., Smith, C., Bonito, A. J. & Steinman, R. M. The linkage of innate to adaptive immunity via maturing dendritic cells in vivo requires CD40 ligation in addition to antigen presentation and CD80/86 costimulation. J. Exp. Med. 199, 1607–1618 (2004).

    CAS  PubMed  PubMed Central  Google Scholar 

  42. Bennett, S. R. et al. Help for cytotoxic-T-cell responses is mediated by CD40 signalling. Nature 393, 478–480 (1998).

    CAS  PubMed  Google Scholar 

  43. Ohshima, Y. et al. OX40 costimulation enhances interleukin-4 (IL-4) expression at priming and promotes the differentiation of naive human CD4+ T cells into high IL-4-producing effectors. Blood 92, 3338–3345 (1998).

    CAS  PubMed  Google Scholar 

  44. Laky, K., Evans, S., Perez-Diez, A. & Fowlkes, B. J. Notch signaling regulates antigen sensitivity of naive CD4+ T cells by tuning co-stimulation. Immunity 42, 80–94 (2015).

    CAS  PubMed  PubMed Central  Google Scholar 

  45. The Tabula Muris Consortium Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris. Nature 562, 367–372 (2018).

  46. Zeisel, A. et al. Molecular architecture of the mouse nervous system. Cell 174, 999–1014 e1022 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  47. Regev, A. et al. The human cell atlas. eLife 6, e27041 (2017).

  48. Kiselev, V. Y., Andrews, T. S. & Hemberg, M. Challenges in unsupervised clustering of single-cell RNA-seq data. Nat. Rev. Genet. 20, 273–282 (2019).

    CAS  PubMed  Google Scholar 

  49. Wolock, S. L., Lopez, R. & Klein, A. M. Scrublet: computational identification of cell doublets in single-cell transcriptomic data. Cell Syst. 8, 281–291 e289 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  50. McGinnis, C. S., Murrow, L. M. & Gartner, Z. J. DoubletFinder: doublet detection in single-cell RNA sequencing data using artificial nearest neighbors. Cell Syst. 8, 329–337 e324 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  51. DePasquale, E. A. et al. DoubletDecon: cell-state aware removal of single-cell RNA-seq doublets. Cell Rep. 29, 1718–1727 (2018).

  52. Krishnaswamy, S. et al. Systems biology. Conditional density-based analysis of T cell signaling in single-cell data. Science 346, 1250689 (2014).

    PubMed  PubMed Central  Google Scholar 

  53. Pasqual, G. et al. Monitoring T cell–dendritic cell interactions in vivo by intercellular enzymatic labelling. Nature 553, 496–500 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  54. Akkaya, B. et al. Regulatory T cells mediate specific suppression by depleting peptide-MHC class II from dendritic cells. Nat. Immunol. 20, 218–231 (2019).

    CAS  PubMed  PubMed Central  Google Scholar 

  55. Yan, J., Liu, B., Shi, Y. & Qi, H. Class II MHC-independent suppressive adhesion of dendritic cells by regulatory T cells in vivo. J. Exp. Med. 214, 319–326 (2017).

    CAS  PubMed  PubMed Central  Google Scholar 

  56. Sakaguchi, S., Yamaguchi, T., Nomura, T. & Ono, M. Regulatory T cells and immune tolerance. Cell 133, 775–787 (2008).

    CAS  PubMed  Google Scholar 

  57. Li, D. et al. VCAM-1+ macrophages guide the homing of HSPCs to a vascular niche. Nature 564, 119–124 (2018).

    CAS  PubMed  PubMed Central  Google Scholar 

  58. Scott, L. M., Priestley, G. V. & Papayannopoulou, T. Deletion of alpha4 integrins from adult hematopoietic cells reveals roles in homeostasis, regeneration, and homing. Mol. Cell Biol. 23, 9349–9360 (2003).

    CAS  PubMed  PubMed Central  Google Scholar 

  59. Deczkowska, A. et al. Disease-associated microglia: a universal immune sensor of neurodegeneration. Cell 173, 1073–1081 (2018).

    CAS  PubMed  Google Scholar 

  60. Camberis, M. et al. Evaluating the in vivo Th2 priming potential among common allergens. J. Immunol. Methods 394, 62–72 (2013).

    CAS  PubMed  Google Scholar 

  61. Paul, F. et al. Transcriptional heterogeneity and lineage commitment in myeloid progenitors. Cell 163, 1663–1677 (2015).

    CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We thank G. Brodsky for artwork. The research of I.A. and A.T. is supported by the Seed Networks for the Human Cell Atlas of the Chan Zuckerberg Initiative and by Merck KGaA, Darmstadt. I.A. is an Eden and Steven Romick Professorial Chair, supported by the HHMI International Scholar Award, the European Research Council Consolidator Grant (no. 724471-HemTree2.0), an MRA Established Investigator Award (no. 509044), DFG (no. SFB/TRR167), the Ernest and Bonnie Beutler Research Program for Excellence in Genomic Medicine, the Helen and Martin Kimmel awards for innovative investigation, and the SCA award of the Wolfson Foundation and Family Charitable Trust. The Thompson Family Foundation Alzheimer’s Research Fund and the Adelis Foundation also provided support. A.T.’s laboratory is supported by the European Research Council, the I-CORE for chromatin and RNA regulation, and a grant from the Israel Science Foundation. A.T. is a Kimmel investigator. A.G. is a recipient of the Clore fellowship. M.C. is supported by a postdoctoral fellowship in Applied and Engineering Science, Israeli Government, Ministry of Science and Technology.

Author information

Authors and Affiliations

Authors

Contributions

A.G. conceived, designed and analyzed experiments; developed the PIC-seq algorithm; performed bioinformatic analysis; and wrote the manuscript. M.C. conceived, designed, performed, and analyzed experiments; developed experimental protocols; and wrote the manuscript. C.M. designed, performed and analyzed experiments. Y.B contributed to PIC-seq algorithm development. B.L., M.Z., T.M.S. and E.D. contributed to the experiments. P.B., R.B. and J.U.M. helped establish the in vivo infection model. F.R. conceived, designed and supervised experiments. A.T. and I.A. directed the project; conceived, designed and analyzed experiments; developed the PIC-seq algorithm. and wrote the manuscript.

Corresponding authors

Correspondence to Amos Tanay or Ido Amit.

Ethics declarations

Competing interests

A patent application has been filed related to this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Integrated supplementary information

Supplementary Figure 1 Data overview.

a, Experimental design of the in vitro model. OT-II CD4+ T cells were co-cultured with ovalbumin (OVA)-loaded, lipopolysaccharides (LPS)-exposed DC for 3, 20 and 44 hours. Mono-cultures of both cell types were used as controls. b, Projection of genes on the Metacell-generated 2d graph (Fig. 1c) allows identification of cell types by expression quantiles of Trbc2 (T cells) and Fscn1 (DC). c, Metacells distribution across MARS-seq technical replicates (rows), grouped by experimental condition and sorting scheme. Color bar indicates assignment of meta-cells to 9 T and DC subsets. d-g, summary of all MARS-seq plates and cells (Supplementary Table 1). Shown are (e) total number of Illumina reads, (f) total number of Unique Molecular Identifiers (UMI) per cell, and (g) fraction of QC-positive cells retained for further analysis per technical replicate. In (d), ‘Replicates’ indicate number of biological replicates (co-cultures or mice), ‘Batches’ indicate number of technical replicates, ‘%analyzed’ indicates fraction of QC-positive cells retained for downstream analysis.

Supplementary Figure 2 PIC-seq performance in a co-culture experiment.

a, Spurious PIC rate is quantified by mixing two parallel cultures, one stained for TCRβ-FITC and CD11c-APC, and the other to TCRβ-PE and CD11c-APC.Cy7. PICs from FITC+APC.Cy7+ and PE+APC+ gates are considered spurious (~1% of all PICs). b, PICs share genes from T and DC. Shown is expression of top 10 differential DC genes (y axis) plotted against top 10 differential T genes (x axis) in single cells (left) and in PICs (right). c, Performance of the linear regression model, used to estimate the mixing factor of 20,000 synthetic PICs. d-e, Performance of the (d) T cell and (e) DC metacell assignments of PIC-seq over 5,000 synthetic PICs. Each row summarizes all synthetic PICs originating from one metacell and their assignments to metacells by PIC-seq (columns. Data is row-normalized). f, Estimation of triplet frequency in PIC-seq data. Triplets can be composed of two T cells and a DC (left), or of one T cell and two DCs (right). Each PIC was assigned a triplet score, indicating the gain in likelihood when instead of modelling doublets, the MLE models heterotypic triplets. Plotted are the cumulative distributions of triplet score of synthetic doublets (gray), triplets (blue), and empirical PIC (green). g, Pearson correlation between genes observed and expected values across all empirical PICs, in a leave 10% out cross-validation. Expected values are calculated (i) from the full PIC-seq algorithm, (ii) when incorporating the meta-cell assignments or (iii) the mixing factor only, or (iv) when mixing factor and metacell assignments are arbitrary. n=2,725 PICs. h, Differential gene expression between singlets from three DC subsets against all other DCs (log2 fold change, x axis), plotted against differential expression between PICs assigned to the same DC subsets and the activated T cell subset, calculated against all other PICs assigned to activated T cells (log2 fold change, y axis).

Source data

Supplementary Figure 3 Activation dynamics of co-cultured T cells and DC.

a-b, Gene modules analysis. Shown are five and four correlated gene modules derived from mono- and co-cultured single T cells (a) and single DCs (b), respectively. Left – pairwise spearman correlation. Right – z-normalized pooled expression across different conditions. n=3,642 T cells (a) and 4,558 DCs (b). c, Pooled expression of five gene modules across single T cells derived from mono-cultures, 3h, 20h and 44h of co-culture. Dots represent single cells and dot colors indicate assignment to T subsets. Representative genes (black) and transcription factors (red) from each module are depicted (Supplementary Table 3). d, Pooled expression of four gene modules across DCs derived from mono-cultures and 3, 20 and 44h of co-culture (Supplementary Table 4). The central mark in box plot is median, with 5/95 percentiles at the whiskers and 25/75 percentiles at the box. Two-tailed Kolmogorov-Smirnov test; n=3,891 DCs. e, Projection of T cells and DCs after 20 hours of mono-culture (top), co-culture (bottom), or co-culture disrupted by an insert preventing cell contact but allowing passage of soluble material (transwell; middle).

Supplementary Figure 4 PIC-seq reveals advanced differentiation in conjugated T cells.

a, Schematics: expected PIC gene expression values are calculated by mixing the multinomial probability vectors of the contributing metacells by the inferred mixing factor, while preserving total UMI count. b, Differences between observed and expected gene expression pooled over all PICs (log2 fold change). Expected values are calculated over the maximum likelihood doublet assignments (x axis), or the maximum likelihood triplet assignments (y axis; of two T cells and one DC; Methods). c, Mean observed and expected gene-expression levels in PIC grouped by their T and DC contributor subsets along the different time points of the co-culture experiment, shown for selected genes. Expected values are decomposed by the contribution from the T (green) and DC (red) PIC components. Bottom panel shows T and DC subset identities. Error bars indicate binomial 95% confidence intervals of the observed values. d, Log2 fold change values of 348 genes exhibiting significant differences between their observed and expected values across PICs grouped by their T and DC contributor subsets. χ2 test; q < 10-6. c-d, n=2,389 PICs.

Supplementary Figure 5 PIC-seq in the postnatal lung.

a, FACS plot of EPCAM+ (CD326) epithelial cells, CD45+ immune cells and PIC derived from postnatal murine lung. Results are representative of two independent experiments. b, Gene expression profiles of 1,071 CD45+ or EPCAM+ single cells from postnatal lungs, grouped into 13 cell types. Top panel indicates UMI count. c, CD45+ and EPCAM+ cell subset identity of single cells in a. d, Gene expression profiles of 543 PICs, grouped by their contributing CD45+ and EPCAM+ identities, as determined by PIC-seq algorithm. Top panels as in a. e, CD45+ and EPCAM+ subset identities of PIC contributing cells in b, as determined using the PIC-seq algorithm. f, Estimation of the relative UMI count from epithelial cells (green), and immune cells (red) contributing to each PIC in b, as inferred using the PIC-seq algorithm (mixing factor). g-h, Differences in metacell composition between PIC contributing epithelial cells (g) and PIC contributing immune cells (h) in postnatal lungs. Each bar represents one metacell, and shows log2 fold change of its frequency between the PIC and the EPCAM+ (g) or CD45+ (h) single-cell populations. Bar colors relate to metacell subset identity. i, Distribution of Epcam+ and CD45+ subsets in non-conjugated populations and in PICs. Two-tailed FDR adjusted Fisher’s exact test; n= 621 Epcam+, 441 CD45+ single cells, and 543 PICs. j, Mean observed (gray bar) and expected (colored bar) gene expression of Alveolar type 1 (AT1)-C45+ PICs grouped by subsets of C45+ contributing cells (x-axis). Error bars indicate binomial 95% confidence intervals of the observed values. n=543 PICs.

Supplementary Figure 6 PIC-seq of an in vivo infection model.

a, Comparison between percentages of singlet T (TCRβ+) and B (CD19+) cells, and T-B PICs achieved by mild and strong dissociation conditions with the enzymes liberase and DNaseI. Results are representative of two tested digestion protocols with similar results. b, Quantification of spurious T-DC PIC rates, by mixing two LNs, one stained for TCRβ-PE and CD11c-APC, and the other to TCRβ-PerCP.Cy7 and CD11c-APC.Cy7. PICs from PE+APC.Cy7+ and PerCP.Cy7+APC+ gates are spurious (summing to <25% of all PICs). Bars indicate mean values; n=2 LNs. c, Full sorting schemes for the Nb immunization experiment. d, Confocal microscopy images of a PIC triplet composed of two T cells and one DC. Scale bar=10μm. Image is representative of four mice from two independent experiments with similar results. e, Metacells distribution across MARS-seq technical replicates (rows), grouped by experimental condition, sorting scheme and biological replicate. Color bar indicates association of meta-cells to 10 cell subsets. f-g, Cell type distributions across different TCRβ+ (f) and CD11c+ (g) sub-sorts and experimental conditions. h, Sorting scheme for Nb-presenting Ag+CD11c+ single cells and Ag+TCRβ+CD11c+ PICs.

Source data

Supplementary Figure 7 Performance of the PIC-seq algorithm on the auricular LN.

a, Performance of the linear regression model on the in vivo infection dataset, used to estimate the mixing factor of 20,000 synthetic PICs. b-c, Performance of the MLE over 5,000 synthetic PICs in T (b) and APC (c) metacell assignments. Each row summarizes all synthetic PICs originating from one metacell, and their assignments to metacells by PIC-seq (columns. Data is row-normalized). d-e, Estimation of triplet frequency in PIC-seq data. Triplets are either composed of two T cells and an APC (d), or one T cell and two APCs (e), as in Supplementary Figure 2. Right panels display performance of triplet filtering method over mixes of doublets with 4% triplets with two T cells (d), or doublets with 3% triplets with two APCs (e). False negative rate (y axis) is plotted against total fraction of discarded cells (x axis). Vertical lines indicate empirical filtering threshold. f, Differential gene expression between singlets from five T subsets against all other T cells (log2 fold change, x axis), plotted against differential expression between PICs assigned to the same T subsets and the migratory DC subset, calculated against all other PICs assigned to migratory DCs (log2 fold change, y axis).

Supplementary Figure 8 A costimulatory module is up-regulated in PIC of antigen presenting DC.

a-b, Cell frequencies of different (a) T cell subsets and (b) APC subsets in the non-conjugated state and in PICs, following PBS injection or Nb infection. Bars depict the mean values across all cells; error bars indicate binomial 95% confidence intervals; lines connect non-conjugated cells and PICs from the same sample and line colors indicate significant differences between single cells and PICs; n=3 (PBS T cells), 4 (Nb T cells), 2 (PBS APCs), 4 (Nb APCs), 3 (PBS PICs) and 5 (Nb PICs) mice; Two-tailed FDR-adjusted Fisher exact test, q < 0.05; for binomial estimation: n=1015 PBS and 1579 Nb T cells, 978 PBS and 1700 Nb DCs, and 977 PBS and 1357 Nb PICs. c, Log2 fold change of 146 genes exhibiting significant differences between their observed and expected values across PICs assigned to Treg, grouped by their assignment to APC subsets. χ2 test; q < 10-6; n=504 PICs d, Frequencies of different T subsets in PICs from Nb infected LN, and from Nb-presenting PIC (Ag+TCRβ+CD11c+). n=Nb-injected PICs from 5 mice, and Nb-presenting PICs from 3 mice. e, Log2 fold change of 216 genes exhibiting significant differences between their observed and expected values across PICs assigned to migratory DC, grouped by whether they were sorted from PBS-injected, Nb-infected (Ag-), or Nb-presenting (Ag+) PIC populations. χ2 test; q < 10-6. f, Mean observed and expected gene-expression levels in PICs grouped as in e, shown for selected genes. Expected values are decomposed by the contribution from the T (green) and DC (red) PIC components. Error bars indicate binomial 95% confidence intervals of the observed values. g, Expression of costimulatory genes across 20h single cells from the coculture experiment. Line colors indicate whether a cell is from the mono-culture, transwell, or co-culture experiments. f-h, n=711 PICs. h, FACS gating scheme for pathogen presenting (Ag+) APCs (CD11c+) and PICs, and non-presenting (Ag-) PICs. i, Histogram of DLL4 expression for the different populations in h. h-i, data is from two mice processed in one experiment.

Supplementary information

Supplementary Materials

Supplementary Figs. 1–8.

Reporting Summary

Supplementary Software

PIC-seq algorithm and auxiliary scripts.

Supplementary Table 1

Summary of all technical replicates.

Supplementary Table 2

Gene lists used for analysis.

Supplementary Table 3

Correlated gene modules across single T cells from in vitro experiment.

Supplementary Table 4

Correlated gene modules across single DCs from in vitro experiment.

Source data

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Giladi, A., Cohen, M., Medaglia, C. et al. Dissecting cellular crosstalk by sequencing physically interacting cells. Nat Biotechnol 38, 629–637 (2020). https://doi.org/10.1038/s41587-020-0442-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41587-020-0442-2

This article is cited by

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing