Detailed characterization of the cell types in the human brain requires scalable experimental approaches to examine multiple aspects of the molecular state of individual cells, as well as computational integration of the data to produce unified cell-state annotations. Here we report improved high-throughput methods for single-nucleus droplet-based sequencing (snDrop-seq) and single-cell transposome hypersensitive site sequencing (scTHS-seq). We used each method to acquire nuclear transcriptomic and DNA accessibility maps for >60,000 single cells from human adult visual cortex, frontal cortex, and cerebellum. Integration of these data revealed regulatory elements and transcription factors that underlie cell-type distinctions, providing a basis for the study of complex processes in the brain, such as genetic programs that coordinate adult remyelination. We also mapped disease-associated risk variants to specific cellular populations, which provided insights into normal and pathogenic cellular processes in the human brain. This integrative multi-omics approach permits more detailed single-cell interrogation of complex organs and tissues.
Subscribe to Journal
Get full journal access for 1 year
only $21.58 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.
Grindberg, R.V. et al. RNA-sequencing from single nuclei. Proc. Natl. Acad. Sci. USA 110, 19802–19807 (2013).
Habib, N. et al. Div-Seq: single-nucleus RNA-seq reveals dynamics of rare adult newborn neurons. Science 353, 925–928 (2016).
Krishnaswami, S.R. et al. Using single nuclei for RNA-seq to capture the transcriptome of postmortem neurons. Nat. Protoc. 11, 499–524 (2016).
Lake, B.B. et al. Neuronal subtypes and diversity revealed by single-nucleus RNA sequencing of the human brain. Science 352, 1586–1590 (2016).
Lacar, B. et al. Nuclear RNA-seq of single neurons reveals molecular signatures of activation. Nat. Commun. 7, 11022 (2016).
ENCODE Project Consortium. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012).
Kundaje, A. et al. Integrative analysis of 111 reference human epigenomes. Nature 518, 317–330 (2015).
Hindorff, L.A. et al. Potential etiologic and functional implications of genome-wide association loci for human diseases and traits. Proc. Natl. Acad. Sci. USA 106, 9362–9367 (2009).
Sos, B.C. et al. Characterization of chromatin accessibility with a transposome hypersensitive sites sequencing (THS-seq) assay. Genome Biol. 17, 20 (2016).
Buenrostro, J.D. et al. Single-cell chromatin accessibility reveals principles of regulatory variation. Nature 523, 486–490 (2015).
Cusanovich, D.A. et al. Multiplex single cell profiling of chromatin accessibility by combinatorial cellular indexing. Science 348, 910–914 (2015).
Jin, W. et al. Genome-wide detection of DNase I hypersensitive sites in single cells and FFPE tissue samples. Nature 528, 142–146 (2015).
Rotem, A. et al. Single-cell ChIP-seq reveals cell subpopulations defined by chromatin state. Nat. Biotechnol. 33, 1165–1172 (2015).
Bushman, D.M. et al. Genomic mosaicism with increased amyloid precursor protein (APP) gene copy number in single neurons from sporadic Alzheimer's disease brains. eLife 4, e05116 (2015).
Gole, J. et al. Massively parallel polymerase cloning and genome sequencing of single cells using nanoliter microwells. Nat. Biotechnol. 31, 1126–1132 (2013).
Klein, A.M. et al. Droplet barcoding for single-cell transcriptomics applied to embryonic stem cells. Cell 161, 1187–1201 (2015).
Macosko, E.Z. et al. Highly parallel genome-wide expression profiling of individual cells using nanoliter droplets. Cell 161, 1202–1214 (2015).
Zheng, G.X. et al. Massively parallel digital transcriptional profiling of single cells. Nat. Commun. 8, 14049 (2017).
Kia, A. et al. Improved genome sequencing using an engineered transposase. BMC Biotechnol. 17, 6 (2017).
Buenrostro, J.D., Giresi, P.G., Zaba, L.C., Chang, H.Y. & Greenleaf, W.J. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat. Methods 10, 1213–1218 (2013).
Corces, M.R. et al. Lineage-specific and single-cell chromatin accessibility charts human hematopoiesis and leukemia evolution. Nat. Genet. 48, 1193–1203 (2016).
Ameur, A. et al. Total RNA sequencing reveals nascent transcription and widespread co-transcriptional splicing in the human brain. Nat. Struct. Mol. Biol. 18, 1435–1440 (2011).
Lake, B.B. et al. A comparative strategy for single-nucleus and single-cell transcriptomes confirms accuracy in predicted cell-type expression from nuclear RNA. Sci. Rep. 7, 6031 (2017).
Heimberg, G., Bhatnagar, R., El-Samad, H. & Thomson, M. Low dimensionality in gene expression data enables the accurate extraction of transcriptional programs from shallow sequencing. Cell Syst. 2, 239–250 (2016).
Zhang, Y. et al. An RNA-sequencing transcriptome and splicing database of glia, neurons, and vascular cells of the cerebral cortex. J. Neurosci. 34, 11929–11947 (2014).
Zhang, Y. et al. Purification and characterization of progenitor and mature human astrocytes reveals transcriptional and functional differences with mouse. Neuron 89, 37–53 (2016).
Tasic, B. et al. Adult mouse cortical cell taxonomy revealed by single cell transcriptomics. Nat. Neurosci. 19, 335–346 (2016).
Darmanis, S. et al. A survey of human brain transcriptome diversity at the single cell level. Proc. Natl. Acad. Sci. USA 112, 7285–7290 (2015).
Zeng, H. et al. Large-scale cellular-resolution gene profiling in human neocortex reveals species-specific molecular signatures. Cell 149, 483–496 (2012).
Buffo, A. & Rossi, F. Origin, lineage and function of cerebellar glia. Prog. Neurobiol. 109, 42–63 (2013).
Saab, A.S. et al. Bergmann glial AMPA receptors are required for fine motor coordination. Science 337, 749–753 (2012).
Butts, T., Green, M.J. & Wingate, R.J. Development of the cerebellum: simple steps to make a 'little brain'. Development 141, 4031–4041 (2014).
Hansen, D.V. et al. Non-epithelial stem cells and cortical interneuron production in the human ganglionic eminences. Nat. Neurosci. 16, 1576–1587 (2013).
Ma, T. et al. Subcortical origins of human and monkey neocortical interneurons. Nat. Neurosci. 16, 1588–1597 (2013).
Mathelier, A. et al. JASPAR 2016: a major expansion and update of the open-access database of transcription factor binding profiles. Nucleic Acids Res. 44, D110–D115 (2016).
Choi, J.W. et al. FTY720 (fingolimod) efficacy in an animal model of multiple sclerosis requires astrocyte sphingosine 1-phosphate receptor 1 (S1P1) modulation. Proc. Natl. Acad. Sci. USA 108, 751–756 (2011).
Groves, A., Kihara, Y. & Chun, J. Fingolimod: direct CNS effects of sphingosine 1-phosphate (S1P) receptor modulation and implications in multiple sclerosis therapy. J. Neurol. Sci. 328, 9–18 (2013).
Angerer, P. et al. destiny: diffusion maps for large-scale single-cell data in R. Bioinformatics 32, 1241–1243 (2016).
Marques, S. et al. Oligodendrocyte heterogeneity in the mouse juvenile and adult central nervous system. Science 352, 1326–1329 (2016).
Gautier, H.O. et al. Neuronal activity regulates remyelination via glutamate signalling to oligodendrocyte progenitors. Nat. Commun. 6, 8518 (2015).
Hines, J.H., Ravanelli, A.M., Schwindt, R., Scott, E.K. & Appel, B. Neuronal activity biases axon selection for myelination in vivo . Nat. Neurosci. 18, 683–689 (2015).
Lundgaard, I. et al. Neuregulin and BDNF induce a switch to NMDA receptor-dependent myelination by oligodendrocytes. PLoS Biol. 11, e1001743 (2013).
Mensch, S. et al. Synaptic vesicle release regulates myelin sheath number of individual oligodendrocytes in vivo . Nat. Neurosci. 18, 628–630 (2015).
Wake, H., Lee, P.R. & Fields, R.D. Control of local protein synthesis and initial events in myelination by action potentials. Science 333, 1647–1651 (2011).
Pozniak, C.D. et al. Sox10 directs neural stem cells toward the oligodendrocyte lineage by decreasing Suppressor of Fused expression. Proc. Natl. Acad. Sci. USA 107, 21795–21800 (2010).
Finzsch, M., Stolt, C.C., Lommes, P. & Wegner, M. Sox9 and Sox10 influence survival and migration of oligodendrocyte precursors in the spinal cord by regulating PDGF receptor alpha expression. Development 135, 637–646 (2008).
Zhao, C. et al. Dual regulatory switch through interactions of Tcf7l2/Tcf4 with stage-specific partners propels oligodendroglial maturation. Nat. Commun. 7, 10883 (2016).
Rocha, H., Sampaio, M., Rocha, R., Fernandes, S. & Leão, M. MEF2C haploinsufficiency syndrome: report of a new MEF2C mutation and review. Eur. J. Med. Genet. 59, 478–482 (2016).
Trynka, G. et al. Chromatin marks identify critical cell types for fine mapping complex trait variants. Nat. Genet. 45, 124–130 (2013).
Ernst, J. et al. Mapping and analysis of chromatin state dynamics in nine human cell types. Nature 473, 43–49 (2011).
Maurano, M.T. et al. Systematic localization of common disease-associated variation in regulatory DNA. Science 337, 1190–1195 (2012).
Zhang, B. et al. Integrated systems approach identifies genetic nodes and networks in late-onset Alzheimer's disease. Cell 153, 707–720 (2013).
Gosselin, D. et al. An environment-dependent transcriptional network specifies human microglia identity. Science 356, eaal3222 (2017).
Fan, H.C., Fu, G.K. & Fodor, S.P. Combinatorial labeling of single cells for gene expression cytometry. Science 347, 1258367 (2015).
Ramani, V. et al. Massively multiplex single-cell Hi-C. Nat. Methods 14, 263–266 (2017).
Zeisel, A. et al. Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq. Science 347, 1138–1142 (2015).
Mattson, M.P. Pathways towards and away from Alzheimer's disease. Nature 430, 631–639 (2004).
Hawrylycz, M.J. et al. An anatomically comprehensive atlas of the adult human brain transcriptome. Nature 489, 391–399 (2012).
Uhlén, M. et al. Tissue-based map of the human proteome. Science 347, 1260419 (2015).
La Manno, G. et al. Molecular diversity of midbrain development in mouse, human, and stem cells. Cell 167, 566–580 (2016).
Yu, G., Wang, L.G. & He, Q.Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382–2383 (2015).
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Flow cytometry was done at both the UCSD Human Embryonic Stem Cell Core and the TSRI Flow Cytometry Core. We thank T.F. Osothprarop and M.M. He (Illumina, San Diego, California, USA) for providing Tn5059 transposase; N. Salathia for assistance in sequencing; Y. Wu, T. Pakozdi, and Z. Chiang for assistance in sequencing analysis; and G. Kennedy for help on RNAscope. Funding support was provided by the NIH Common Fund Single Cell Analysis Program (1U01MH098977 to K.Z. and J.C.). T.E.D. is a fellow in the Pediatric Scientist Development Program and is supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (award no. K12-HD000850). G.E.K. was supported by a Neuroplasticity of Aging Training Grant (5T32AG000216-24). P.V.K. was supported by the NIH Centers for Excellence in Genomic Science (P50MH106933), NIH grant 1R01HL131768, and an NSF CAREER award (NSF-14-532). J.F. was supported by NIH grant F31 CA206236. The Genotype-Tissue Expression (GTEx) Project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by NCI, NHGRI, NHLBI, NIDA, NIMH, and NINDS. The GTEx samples used for the analyses described in this study were obtained from the University of Miami Brain Endowment Bank.
The authors declare no competing financial interests.
Integrated supplementary information
A. snDrop-seq method showing modifications needed to process nuclei, including bovine serum albumin (BSA) coating and droplet heating to ensure complete nuclear membrane lysis. B. Violin plot showing the ratio of area size of DAPI staining versus droplet size of untreated and heat-treated samples, confirming efficient nuclear lysis in snDrop-seq. Top left panel: 293T cells pre-stained with DAPI showing cellular, but not nuclear lysis inside droplets. Top right panel: DAPI staining in droplets after applying heat-treatment showing efficient lysis as evidenced by the overspreading of DNA stain inside the droplets. C. Collision plot for human (293T) and mouse (NIH3T3) nuclei analyzed by a single snDrop-seq run. Scatter plot shows the number of transcript reads mapping to human (hg19) and mouse (mm10) genomes for each cell barcode. Percent of collisions is indicated. D. Gene body coverage for visual cortex data sets (library occ5). E. Proportion of reads mapped to exons, 5’ untranslated regions (UTR) 3’ UTRs, introns, intergenic regions and that are unassigned across quality filtered data sets from the visual cortex (occ1-21, Table S1). F. Number of reads associated with (E).
A. For combinatorial barcoding, ~2000 nuclei are added per well to a 384-well plate containing uniquely barcoded transposons, followed by tagmentation, inactivation, pooling, and FACS redistribution into 96-well plates. In vitro transcription (IVT) and scTHS-seq RNA-seq is performed followed by PCR indexing, pooling and high throughput sequencing of libraries. Optimally ~25,000 total single-cell datasets are recovered per run. B. Customized scTHS-seq transposon design. Transposons contain a mosaic end for Tn5 binding, read primer site for sequencing, an r5 barcode unique to every transposon, an i5 connector sequence used in PCR amplification and a T7 promoter for IVT amplification. To generate the barcoded transposome complexes, each transposon is annealed to the RC mosaic end, then transposase is added. C. Depiction of sequencing ready fragments generated after library generation. Fragments contain i5/i7 adaptors and barcodes, the r5 transposon barcode, Read1 sequencing primer, and captured genomic DNA. The i5 and r5 barcodes are read in one sequencing read, with sequencing progressing through the linker region and ending with the r5 barcode. D. Confirmation of in vitro transcription and barcoded library generation in scTHS-seq. Left panel: gel image of whole reactions purified with SPRI binding buffer. Occipital lobe sample plate 5 and plate 6 quality controls (QC) consisted of two 100 nuclei samples, and two no template controls (NTCs). Right panel: gel images of occipital lobe samples from 12 wells from a 96-well plate, showing library size range that is selected for after pooling all samples. E. Identification of cell type specific DNA accessibility peaks over the promoter and regulatory region of Alzheimer’s disease associated gene PICALM. Glial specific peaks are over-represented. The top tract shows all cells merged to generate peaks. Each cell subpopulation tract is represented by 100 randomly selected single cells having reads in the depicted region, where each row represents a single cell and each dot is a read. The color of highlighted peak regions corresponds to the cell type specificity of the peak, with each subpopulation tract title a specific color. Boxed regions highlight specific cell type specific peaks. The gray box highlights a glial cell specific peak. Nonspecific peaks are not highlighted.
A. Cluster dendrogram showing unsupervised grouping of all non-neuronal, excitatory and inhibitory neuronal cell types or subtypes across cortical and cerebellar regions. B. t-SNE plot for all data sets as shown in Fig. 1B. C. t-SNE plot as in (B) showing batch identity (see Table S1). D. t-SNE plot as in (B) showing individual patient identity (see Table S1). E. t-SNE plot as in (B) showing UMI read depth. Non-neuronal cell types show lower coverage than neuronal subtypes. F-I. Same as for (B-E) except showing only visual cortex (Occ or occipital lobe, BA17) data. Some patient-specific clustering can be seen for Ex3 neurons that may be associated with different visual cortical regions sampled between repositories. J-M. Same as for (B-E) except showing only frontal cortex (FCtx, BA6, BA8) data. N-Q. Same as for (B-E) except showing only cerebellar hemisphere (CBL) data.
A. snDrop-seq clusters from the human brain visualized by t-SNE (from Fig. 1B). B. Mouse retina data generated from single cells (sc) using the Drop-seq method and previously annotated17 were visualized using a maximum of 100 cells per cluster by t-SNE. C. Mouse embryonic brain (E18) data sets generated on the 10X Genomics platform were clustered and visualized by t-SNE. D. Embryonic human midbrain data sets generated on the Fluidigm C1 platform and previously annotated61 were visualized by t-SNE. E-H. Histograms of detected molecules (number of UMI) for associated quality filtered data sets. I-L. Histograms of detected genes for associated quality filtered data sets. M-P. Coverage (number of UMI per gene detected) for associated quality filtered data sets. Q. Scatter plot showing average expression values for protein-coding genes across quality filtered snDrop-seq samples (occ1-21, log transformed) and their associated genic length. R. Scatter plot showing average expression values for protein-coding genes across all quality filtered scDrop-seq samples (log transformed) and their associated genic length.
A-B. Frequency of unique reads in each possible barcode combination in the human (hg38 noAlt) and mouse (mm10 noAlt) datasets. C-D. Frequency of percent clonal reads in each possible barcode combination in the human (hg38 noAlt) and mouse (mm10 noAlt) datasets. Clonal read statistics were rounded to the nearest whole percent before calculating. E. Collision density plot showing the ratio of unique reads mapped to human and unique reads mapped to mouse in each unique barcode combination (data with <500 reads were excluded). See Methods for determination of which data sets are considered collisions. F. Theoretical and experimentally derived collision rate percentages. Note that collision rates cannot be directly measured for mouse or human only nuclei. G. Violin plot of unique read counts for nuclei that pass filter and nuclei that pass filter with multi-mappers removed. The blue dots represent average number of unique reads and the red dots represent the median number of unique reads. H. Violin plot as in (G) showing only unique read count range of 0 to 80,000. I. Visualization of merged single cell data and individual single cell datasets from the visual cortex. All unique reads that passed filter were combined to generate the merged tract. Single cell datasets that had reads present in the chromosome 17 locus from 8,100,000 to 8,250,000 bp were randomly sampled and plotted in the single cell tracts section. Each black dot represents a unique read. UCSC genes are overlaid at the bottom.
A. t-SNE plot for all data sets. Cluster dendrogram showing unsupervised grouping of all non-neuronal, excitatory and inhibitory neuronal cell types or subtypes across cortical and cerebellar regions (top right). t-SNE plot showing batch identity (bottom left). t-SNE plot showing read depth (bottom right). B. Same as for A except showing only visual cortex (Occ or occipital cortex, BA17) data. C. Same as for A except showing only frontal cortex. D. Same as for A except showing only cerebellum. For cluster proportions see Table S2.
A. Correlation heatmaps comparing averaged snDrop-seq data from all regions sampled with average expression values from RNA-seq data from: mouse pooled cortical cell types25 (left, nf = newly formed); human pooled temporal lobe cell types26 (right). B. Correlation heatmaps comparing average visual cortex data with average expression values from single-cell RNA-seq data from the mouse visual cortex27 (left) and human temporal lobe28 (right). C. Correlation heatmap of snDrop-seq-identified neuronal subtypes (visual cortex or frontal cortex) compared with subtypes previously identified using the C1 single nucleus Smart-seq+ pipeline4 (SNS, across cortical regions). D. Pair-wise correlation values for all cell types and subtypes resolved across cortical and cerebellar brain regions. E. Cell type sampling rates from single-cell RNA-seq on the temporal lobe28 compared with snDrop-seq sampling rates from the visual and frontal cortices.
A. Heatmap of top 10 differentially expressed marker genes enriched across excitatory neuron subtypes (Table S3). B. Heatmap of the top 10 differentially expressed marker genes enriched across inhibitory neuron subtypes (Table S3). C. Heatmap of the top 10 differentially expressed marker genes enriched across cerebellar clusters (Table S3). D. Top panel: RNA ISH counts showing number of positive cells for CBLN2 and PCP4 (chromogenic image shown) in image fields spanning the pial layer to the white matter. Lower panel: RNA ISH counts for SLC17A7 and EYA4 single positive cells. E. RNA ISH stains (Table S9) of the visual cortex showing select markers from Fig. 2C, and predicted spatial distribution of associated In neuron subtypes.
A. Schematic of the cerebellar hemisphere cytoarchitecture as shown in Fig. 2E. B. Violin plots of expression values for type-specific marker genes specifically for cerebellar data. C. Protein staining (Table S10) for select cell-type specific markers shown in (B). Arrows indicate positively stained cells. D. Representative RNA ISH stainings for GAD1+SORCS3+ (left) and GAD1+SORCS3− (right) Purk neurons that were quantified in Fig. 2H.
Supplementary Figure 10 Probabilistic inference of differential expression and differential accessibility signals.
A-C. Predicting differential accessibility of sites from differential expression. Astrocytes (Ast) and Oligodendrocytes (Oli) identified independently from snDrop-seq and scTHS-seq data analysis are used for training and testing. Genes highly upregulated in Ast snDrop-seq population predict sites highly accessibility in the Ast scTHS-seq population and similarly for Oli. A. A receiver operating characteristic (ROC) curve is used to illustrate the accuracy of predicting the differentially accessible sites within the proximity of the differentially-expressed genes. B. The ROC curve shows the performance of cell classification based on a multi-site predicted differential accessibility signature. In this case, we assess the ability of the average expected differential accessibility inferred from the differential expression signature to be able to distinguish Ast and Oli groups in the scTHS-seq data. C. Predicted differentially accessible sites in Ast and Oli are visualized with red indicating higher aggregate accessibility of oligodendrocyte-associated genes and blue for astrocyte. True cell population assignments are labeled. D-F. Predicting differential expression from differential accessibility profiles. D. ROC of predicting differential expression of an individual gene is shown. E. ROC curve and visualization of the combined differential expression scores inferred from the differential accessibility signature comparing Ast and Oli groups. F. Predicted differentially expressed genes in Ast and Oli are visualized with red indicating higher aggregate expression of oligodendrocyte-associated genes and blue for astrocyte. True cell population assignments are labeled.
Supplementary Figure 11 Transcription factors associated with cell-type-specific chromatin accessibility tend to be highly expressed in the associated cell type.
A. Oli-associated TFs are significantly upregulated in Oli based on the fold-change comparing TF expression in oligodendrocytes to neurons. B. Oli-associated TFs are significantly depleted in neurons based on the fold-change comparing neurons to oligodendrocytes.
A. 3D visualization of Destiny components 1, 2, and 3 from visual cortex scTHS-seq data for OPCs and Oli only. Cell population assignments from the original clustering analysis are shown in blue for OPCs and red for Oli. B. 2D visualization of Destiny components 1, 2. C-D. Same plots as A-B colored by batch. E. Heatmap of all significantly (Z > 1.96) differentially upregulated genes corresponding to OPC, iOli and mOli.
A. Single-cell ordering OPC and Oli data along a remyelination trajectory was generated using Monocle (see Methods). Top plot shows cluster identities (OPC, OPC_Cer and Oli), bottom plot shows regional identities (Cerebellum or CBL, Frontal Cortex and Visual Cortex). B. Pseudotemporal heatmap showing gene expression modules (I to VI), ordered based on expression dynamics progressing along the predicted remyelination trajectory, and generated from gene sets shown in Fig. S12E. Top 5 gene ontologies for each module are indicated.
Supplementary Figure 14 Mapping of disease risk variants to specific brain cell types in four brain diseases and seven non-brain diseases.
Z-scores for the enrichment of GWAS SNPs in the open chromatin were overlaid onto the Ex, In, Oli, OPC, Ast, End, Mic cell populations (Table S8). Three autoimmune diseases (Crohn’s disease, Celiac disease, T1D) were included. Dark purple and purple represent a significant Z-score over 1.96, where light purple, gray and light green represent an insignificant Z-score, and green represents a significant negative association with a Z-score less than -1.96.
About this article
Cite this article
Lake, B., Chen, S., Sos, B. et al. Integrative single-cell analysis of transcriptional and epigenetic states in the human adult brain. Nat Biotechnol 36, 70–80 (2018). https://doi.org/10.1038/nbt.4038
Single-Nuclei RNA Sequencing Assessment of the Hepatic Effects of 2,3,7,8-Tetrachlorodibenzo-p-dioxin
Cellular and Molecular Gastroenterology and Hepatology (2021)
Biological Psychiatry (2020)
Cancer Cell (2020)
Bystander IFN-γ activity promotes widespread and sustained cytokine signaling altering the tumor microenvironment
Nature Cancer (2020)
Experimental & Molecular Medicine (2020)