Single-cell transcriptome analysis of lineage diversity in high-grade glioma - PubMed (original) (raw)

doi: 10.1186/s13073-018-0567-9.

Hanna Mendes Levitin 1, Veronique Frattini 2, Erin C Bush 1 3, Deborah M Boyett 4, Jorge Samanamud 4, Michele Ceccarelli 5, Athanassios Dovas 6, George Zanazzi 6, Peter Canoll 6, Jeffrey N Bruce 4, Anna Lasorella 2 6 7, Antonio Iavarone 2 6 8, Peter A Sims 9 10 11

Affiliations

Single-cell transcriptome analysis of lineage diversity in high-grade glioma

Jinzhou Yuan et al. Genome Med. 2018.

Abstract

Background: Despite extensive molecular characterization, we lack a comprehensive understanding of lineage identity, differentiation, and proliferation in high-grade gliomas (HGGs).

Methods: We sampled the cellular milieu of HGGs by profiling dissociated human surgical specimens with a high-density microwell system for massively parallel single-cell RNA-Seq. We analyzed the resulting profiles to identify subpopulations of both HGG and microenvironmental cells and applied graph-based methods to infer structural features of the malignantly transformed populations.

Results: While HGG cells can resemble glia or even immature neurons and form branched lineage structures, mesenchymal transformation results in unstructured populations. Glioma cells in a subset of mesenchymal tumors lose their neural lineage identity, express inflammatory genes, and co-exist with marked myeloid infiltration, reminiscent of molecular interactions between glioma and immune cells established in animal models. Additionally, we discovered a tight coupling between lineage resemblance and proliferation among malignantly transformed cells. Glioma cells that resemble oligodendrocyte progenitors, which proliferate in the brain, are often found in the cell cycle. Conversely, glioma cells that resemble astrocytes, neuroblasts, and oligodendrocytes, which are non-proliferative in the brain, are generally non-cycling in tumors.

Conclusions: These studies reveal a relationship between cellular identity and proliferation in HGG and distinct population structures that reflects the extent of neural and non-neural lineage resemblance among malignantly transformed cells.

PubMed Disclaimer

Conflict of interest statement

Tissue was procured from de-identified patients who provided written informed consent to participate in these studies through a protocol approved by the Columbia Institutional Review Board (IRB-AAAJ6163). Research was conducted in accordance with the principles of the Declaration of Helsinki.

Competing interests

Columbia University has filed patent applications based on the technology used in these studies with JY and PAS included as inventors. The remaining authors declare that they have no competing interests.

Publisher’s Note

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

Figures

Fig. 1

Fig. 1

a t-SNE projections of scRNA-Seq profiles for each tumor colored by unsupervised clustering resulting from Phenograph analysis. We note that while the putatively transformed populations in each tumor appear in red for simplicity, the majority of them actually contain multiple Phenograph clusters as shown in d and detailed in Fig. 3. The cell type labels are based on marker expression patterns shown in Additional file 1: Figures S2–9. b Principal component analysis of the z-scored matrix of average chromosomal expression for each tumor showing a characteristic axis of variation, which we call the “malignancy score”, on which the putatively transformed cells are separated from the untransformed cells in each tumor. c Same as a but colored based on the malignancy score in b. d Distributions of malignancy scores for each Phenograph cluster in a showing that all of the putatively transformed clusters have higher median scores than all of the untransformed clusters within each tumor. Stars indicate the putatively transformed clusters. e Heatmaps showing the average copy number of each chromosome based on low-pass, bulk WGS (top) and heatmaps showing the average expression of each chromosome in each cell associated with a transformed cluster relative to the average untransformed cell in each tumor (bottom). The high resolution version of Figure 1 is also available as Additional file 2

Fig. 2

Fig. 2

a Analysis of the pervasiveness of genes that are highly specific to the transformed cells across all eight patients based on differential expression analysis (see “Methods”; all genes displayed have eightfold specificity for the transformed cells). The colorbar represents the product of the _x_- and _y_-axes. SOX2 is the most pervasively detected gene specific to transformed glioma cells in these eight HGG patients. b Drop-out curve for the total population of transformed cells showing the characteristic sigmoidal shape that indicates how, for the majority of genes, higher expression (counts per thousand or CPT) leads to detection in a higher fraction of cells. Because the detection frequency of SOX2 is close to that of similarly expressed genes, SOX2 is unlikely to be associated with a specific subpopulation of transformed cells and the frequency with which it is expressed among transformed cells is likely to be underestimated by our data. c IHC analysis confirming widespread protein expression of SOX2 in tissue slices from the six of the eight HGG patients in our cohort from which tissue was available for staining. We note that a considerable fraction of unstained nuclei in these specimens appear to be associated with blood vessels

Fig. 3

Fig. 3

a t-SNE projections of the transformed population of cells from each of the eight HGGs from scRNA-Seq. The projections are colored based on the cellular subpopulations identified from unsupervised clustering. b Heatmaps showing the detection frequency of canonical astrocyte, OPC, oligodendrocyte, and neuroblast markers found to be specifically associated with transformed cellular subpopulations shown in a across multiple patients along with SOX2, which is expressed across all transformed populations. The orange heatmap below each green heatmap shows the average detection frequency of cell cycle control genes found in each subpopulation. Note that some tumors have subpopulations resembling multiple neural lineages (PJ016, PJ018, PJ030, PJ048), while others exhibit a relative loss of neural lineage identity and concomitant reduction in proliferation

Fig. 4

Fig. 4

Force-directed graphs generated from the _k_-nearest neighbor graphs of the transformed cells profiled in each patient. Colors indicate which of the astrocyte marker GFAP, the OPC marker OLIG1, the oligodendrocyte marker MOG, or the neuroblast marker STMN2 is most highly expressed in a given cell. For example, a purple cell has higher levels of STMN2 than the other three markers. None of the four markers are detected in white-colored cells. PJ016, PJ018, and PJ048 form multi-branching structures associated specific neural lineages and their respective single-cell average profiles closely resemble the proneural subtype of GBM. For example, one branch of PJ018 terminates with GFAP-expressing astrocytic cells, whereas the other resembles oligodendrocyte differentiation. PJ030, PJ025, and PJ035 are somewhat less structured (although PJ030 contains clearly separated OPC- and astrocyte-like branches) and have single-cell average profiles that closely resemble the classical subtype of GBM. In contrast, PJ017 and PJ032 are unstructured, do not exhibit branching, show reduced neural lineage diversity, and have single-cell average profiles that closely resemble the mesenchymal subtype of GBM

Fig. 5

Fig. 5

a Hierarchical clustering of the correlation between each transformed subpopulation and a database of cell type-specific expression profiles with high variability across the data set. We find three cell type clusters referred to as Neural/ESC, Immune, and Mesenchymal/MSC which divide the tumor cell subpopulations into three major groups. b Gene ontology analysis of the differentially expressed genes between the group III tumors (PJ017/PJ032) and the remaining tumors (PJ016, PJ018, PJ025, PJ030, PJ035, PJ048) after removal of genes specific to the untransformed immune cells in PJ017 and PJ032. The group III tumors show a clear immunological gene signature that is specific to the transformed cells

Fig. 6

Fig. 6

t-SNE projections of scRNA-Seq profiles from all eight tumors. The plots are colored by expression of either CSF1, a macrophage stimulating cytokine, or the gene encoding its cognate receptor CSF1R. Receptor expression is widespread among myeloid cells, but expression of the cytokine is significantly higher in the transformed glioma cells of PJ017 and PJ032 than in the other tumors. We note that no myeloid cells were detected in PJ016

Similar articles

Cited by

References

    1. Huse JT, Holland EC. Targeting brain cancer: advances in the molecular pathology of malignant glioma and medulloblastoma. Nat Rev Cancer. 2010;10(5):319–331. doi: 10.1038/nrc2818. - DOI - PubMed
    1. Sottoriva A, Spiteri I, Piccirillo SG, Touloumis A, Collins VP, Marioni JC, et al. Intratumor heterogeneity in human glioblastoma reflects cancer evolutionary dynamics. Proc Natl Acad Sci U S A. 2013;110(10):4009–4014. doi: 10.1073/pnas.1219747110. - DOI - PMC - PubMed
    1. Gill BJ, Pisapia DJ, Malone HR, Goldstein H, Lei L, Sonabend A, et al. MRI-localized biopsies reveal subtype-specific differences in molecular and cellular composition at the margins of glioblastoma. Proc Natl Acad Sci U S A. 2014;111(34):12550–12555. doi: 10.1073/pnas.1405839111. - DOI - PMC - PubMed
    1. Patel AP, Tirosh I, Trombetta JJ, Shalek AK, Gillespie SM, Wakimoto H, et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Science. 2014;344(6190):1396–1401. doi: 10.1126/science.1254257. - DOI - PMC - PubMed
    1. Szerlip NJ, Pedraza A, Chakravarty D, Azim M, McGuire J, Fang Y, et al. Intratumoral heterogeneity of receptor tyrosine kinases EGFR and PDGFRA amplification in glioblastoma defines subpopulations with distinct growth factor response. Proc Natl Acad Sci U S A. 2012;109(8):3041–3046. doi: 10.1073/pnas.1114033109. - DOI - PMC - PubMed

Publication types

MeSH terms

Grants and funding

LinkOut - more resources