Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Dec;68(12):2486-2502.
doi: 10.1002/glia.23866. Epub 2020 Jul 4.

Multimodal single-cell analysis reveals distinct radioresistant stem-like and progenitor cell populations in murine glioma

Affiliations

Multimodal single-cell analysis reveals distinct radioresistant stem-like and progenitor cell populations in murine glioma

Jes Alexander et al. Glia. 2020 Dec.

Abstract

Radiation therapy is part of the standard of care for gliomas and kills a subset of tumor cells, while also altering the tumor microenvironment. Tumor cells with stem-like properties preferentially survive radiation and give rise to glioma recurrence. Various techniques for enriching and quantifying cells with stem-like properties have been used, including the fluorescence activated cell sorting (FACS)-based side population (SP) assay, which is a functional assay that enriches for stem-like tumor cells. In these analyses, mouse models of glioma have been used to understand the biology of this disease and therapeutic responses, including the radiation response. We present combined SP analysis and single-cell RNA sequencing of genetically-engineered mouse models of glioma to show a time course of cellular response to radiation. We identify and characterize two distinct tumor cell populations that are inherently radioresistant and also distinct effects of radiation on immune cell populations within the tumor microenvironment.

Keywords: SP analysis; glioma; glioma stem cells; myeloid cells; radiation response; radioresistance; single-cell RNA sequencing.

PubMed Disclaimer

Figures

FIGURE 1
FIGURE 1
Side population (SP) analysis of a PDGF‐driven mouse model of glioblastoma at early time points after radiation shows that stem‐like cells of the SP are relatively radioresistant and are enriched. (a) A Kaplan–Meier plot showing that mice treated with 10 Gy lived a median of 20 days longer than untreated mice. (b) SP analysis, using Hoechst 33342 dye exclusion assay, of tumors from mice harvested at symptom onset or 8 or 72 hr after IR shows a higher percentage of cells in the SP at 72 hr as compared to control. Error bars represent the SE of the mean. (c) Representative flow cytometry plots, as quantified in (b). SP cells are poorly stained by Hoechst dyes due to efflux pump dye removal, whereas the main population (MP) is highly stained. The percentage of SP cells is highest at 72 hr after IR, but returns to the same level as the control at recurrence. Insets show treatment prior to SP analysis with verapamil as a control, which inhibits the efflux pump, abrogates the SP, and confirms the SP analysis gating strategy. (d) Flow cytometry plots of tumors without and at 72 hr after IR showing SP analysis of Olig2‐expressing tumor cells (GFP+) and tumor cells derived from the earliest tumor cells (RFP+). SP analysis of all tumor cells (Total) are shown prior to gating for GFP positivity (GFP only, top row) or RFP positivity (RFP only, bottom row). GFP+ cells are exclusively in the MP without and at 72 hr after IR, and RFP+ cells are heavily enriched at 72 hr after IR in the SP as compared to control
FIGURE 2
FIGURE 2
Single‐cell RNA sequencing (scRNA‐seq) identifies tumor and normal brain cell‐types, including normal cells with tumor‐like expression patterns. (a) UMAP embedding and Louvain clustering of scRNA‐seq data including all normal brain and tumor samples across all treatment groups (no IR and 8 ± 2 and 72 hr after 10 Gy) colored by cell‐types. Each dot represents a cell. (b) UMAP plots with each cluster colored by the fraction of cells in the cluster expressing common cell‐type‐specific markers or the RCAS viral transcript confirm cell‐type assignments. (c) Separate UMAP plots of brain and tumor samples show much higher representation of the OPC/tumor cluster in the tumor samples confirming that this cluster includes tumor. Some tumor‐like normal cells are also present in normal brain. (d) Louvain clustering shows multiple expression profiles within the OPCs and Tumor supercluster (left). Clusters are numbered for easier reference. RCAS viral transcript expression varies in this supercluster being lowest in the “OPC & Oligo (Immature)” cluster (Cluster 17) and highest in the OPC/tumor Cluster 14 (right)
FIGURE 3
FIGURE 3
Tumor cell expression at 8 hr after IR shifts toward genes associated with the DNA damage signaling response and cell cycle arrest, followed by cell death leaving two radioresistant groups of cells (Clusters 17 and 23 and Cluster 22) at 72 hr after IR. (a) UMAP plots of the OPC/tumor supercluster by treatment group (No IR and 8 and 72 hr after IR) and sample type (brain and tumor) shows little change in normal brain samples, but a left‐to‐right shift between No IR and 8 hr after IR samples followed by a loss of most tumor cells at 72 hr after IR. Two groups of cells (Clusters 17 and 23 and Cluster 22) remain. Cluster numbers are shown in the upper left panel. (b) UMAP plots of samples without and 8 hr after IR showing the tumor supercluster with a best line drawn by support vector machine (SVM) to divide the cells into right (more similar to 8 hr after IR) and left (more similar to No IR) sides of the shift. (c) GO terms for enriched gene sets, as determined by Gene Set Enrichment Analysis (GSEA) comparing cells on the right of the left‐to‐right shift to those on the left, are associated with processes that are expected after IR. (d) Transcripts per cell of Mki67, Top2a, Hmga2, and RCAS in the OPC/tumor supercluster without (No IR) and 8 and 72 hr after IR are shown. The green to red spectrum represents cells with at least 1 transcript (see scales). Gray represents cells without any transcripts. Rapidly proliferating cells (Cluster 22 expressing Mki67 and Top2a) and slowly proliferating cells (Clusters 17 and 23) are enriched at 72 hr after IR, whereas Hmga2 expressing cells are enriched at 8 hr after IR and then depleted by 72 hr. RCAS viral transcript expression is similar in all treatment groups
FIGURE 4
FIGURE 4
Clusters 17 and 23 belong to the cell lineage leading to mature oligodendrocytes, while Cluster 22 has characteristics of stemness and proliferation. (a) UMAP re‐embedding of Clusters 17, 23, and the mature oligodendrocyte cluster outside the context of all other cells (top panel) shows a continuum of expression between these clusters. The fractions of cells in each cluster expressing cell‐type specific markers (bottom panels) confirm the cell‐types represented by each cluster. The smaller cluster likely represents collisions or doublets. (b) Dot plot showing NES and −Log(q‐value) of master regulators for Cluster 22, including Tead1, Smad3, and Fosl2 (red), which are activated downstream of Hippo signaling and promote oncogenesis, stemness, proliferation, and radioresistance
FIGURE 5
FIGURE 5
Immunohistochemistry and immunofluorescence microscopy show loss of the bulk of tumor cells, which express OLIG2, but survival of the Cluster 22‐like, YAP1‐expressing, proliferative cells of the perivascular niche at 72 hr after IR. (a) Hematoxylin and eosin (H&E) staining and immunohistochemistry with antibodies to OLIG2, Ki67, and Nestin were performed on tumors without IR (no IR) and at 72 hr and at recurrence after IR. Tumor density and the percentage of OLIG2‐expressing cells, which make up the bulk of tumor cells, are reduced significantly at 72 hr after IR, but return to baseline at recurrence, that is, levels similar to unirradiated tumor. The percentage of Ki67‐expressing cells is also reduced at 72 hr after IR, and those that remain are located in the perivascular niche. Ki67‐expressing cells also return to baseline at recurrence. Nestin‐expressing cells, which are located within the perivascular niche, remain at similar levels between treatment conditions. (b) Immunofluorescence microscopy using antibodies to OLIG2 (white), Ki67 (green), and YAP1 (red) was performed on tumor without and at 72 hr after IR (two examples shown). The top row (Merge + DNA) shows an overlay of the staining by all three antibodies along with DAPI to indicate DNA (blue). At 72 hr after IR, the rare cells expressing Ki67 also express YAP1, which are both specific to Cluster 22, while the OLIG2 expressing cells, which make up the tumor bulk, are distinct from those expressing either Ki67 or YAP1 and are mostly lost
FIGURE 6
FIGURE 6
Subclustering of the myeloid cell supercluster reveals four clusters of microglia and bone marrow‐derived macrophages (BMDM) that vary in percentage by sample type and treatment group, some of which have phagocytosed tumor cells. (a) UMAP re‐embedding and Louvain clustering of the myeloid cell supercluster outside the context of all other cells reveals four clusters: Cluster A is tumor‐associated, activated microglia and BMDM that have not phagocytosed tumor cells; Cluster B is non‐tumor‐associated, resting microglia; Cluster C is tumor‐associated microglia and BMDM that have phagocytosed tumor cells; and cluster D is non‐tumor‐associated microglia that have phagocytosed neural cells. (b) UMAP plots with each cluster colored by the fraction of cells in the cluster expressing common cell‐type‐specific markers or the RCAS viral transcript. (c) Fragments per kilobase of transcript per million mapped reads (FPKM) from bulk RNA‐seq of sorted naïve microglia (Naïve MG), blood monocytes (Naïve Mo), tumor‐associated microglia (Tumor MG), and tumor‐associated BMDM (Tumor MF) for common cell type‐specific markers. Combining data from (b) and (c) allows assignment of cell‐types to clusters in (a). (d) Percentages of clusters by treatment group (No IR and 8 and 72 hr after IR) and sample type (brain and tumor) are shown. In normal brain, the percentage of resting microglia (Cluster B), is mostly unchanged between treatment groups. In tumor without IR, the percentage of tumor‐associated, activated microglia and BMDM (Cluster C) is elevated as compared to normal brain. The percentage of tumor‐associated microglia and BMDM (Clusters A and C) increase at the expense of the resting microglia at 8 hr after IR. At 72 hr, the tumor‐associated microglia and BMDM that have not phagocytosed tumor cells (Cluster A) increase further, but those that have phagocytosed tumor cells (Cluster C) decrease consistent with loss of most of the tumor cells at this time. (e) Flow cytometry using cell surface markers (Figure S6 and Methods) was performed on brain with tumor to examine RFP‐positivity in microglia (CD11b+CD45LoLy6gLy6c) and BMDM (CD11b+CD45Hi Ly6gLy6c) without IR and at 8 and 72 hr after IR. RFP‐positivity indicates cells that had phagocytosed RFP‐expressing tumor cells. RFP+ cells are mostly BMDM without IR, but are mostly microglia at 8 hr after IR. (f) Percentages of myeloid cells (CD45+CD11b+) that are RFP+ without IR and at 8 and 72 hr after IR in brain with tumor. Error bars represent the standard deviation. RFP+ cells peak at 8 hr after IR and mostly disappear by 72 hr. These data are consistent with Cluster C in (a)

References

    1. Badri, H. , Pitter, K. , Holland, E. C. , Michor, F. , & Leder, K. (2016). Optimization of radiation dosing schedules for proneural glioblastoma. Journal of Mathematical Biology, 72(5), 1301–1336. 10.1007/s00285-015-0908-x - DOI - PubMed
    1. Bao, S. , Wu, Q. , McLendon, R. E. , Hao, Y. , Shi, Q. , Hjelmeland, A. B. , … Rich, J. N. (2006). Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature, 444(7120), 756–760. - PubMed
    1. Bennett, M. L. , Chris Bennett, F. , Liddelow, S. A. , Ajami, B. , Zamanian, J. L. , Fernhoff, N. B. , … Barres, B. A. (2016). New tools for studying microglia in the mouse and human CNS. Proceedings of the National Academy of Sciences, 113(12), E1738–E1746. 10.1073/pnas.1525528113 - DOI - PMC - PubMed
    1. Bhat, K. P. L. , Salazar, K. L. , Balasubramaniyan, V. , Wani, K. , Heathcock, L. , Hollingsworth, F. , … Aldape, K. D. (2011). The transcriptional coactivator TAZ regulates mesenchymal differentiation in malignant glioma. Genes and Development, 25(24), 2594–2609. - PMC - PubMed
    1. Bleau, A.‐M. , Hambardzumyan, D. , Ozawa, T. , Fomchenko, E. I. , Huse, J. T. , Brennan, C. W. , & Holland, E. C. (2009). PTEN/PI3K/Akt pathway regulates the side population phenotype and ABCG2 activity in glioma tumor stem‐like cells. Cell Stem Cell, 4(3), 226–235. - PMC - PubMed

Publication types