Published OnlineFirst November 29, 2019; DOI: 10.1158/0008-5472.CAN-18-2932 CANCER RESEARCH | TUMOR BIOLOGY AND IMMUNOLOGY Single-Cell Gene Expression Analyses Reveal Distinct Self-Renewing and Proliferating Subsets in the Leukemia Stem Cell Compartment in Acute Myeloid Leukemia Karen Sachs1,2, Aaron L. Sarver3, Klara E. Noble-Orcutt1,3, Rebecca S. LaRue1,3, Marie Lue Antony1,3, Daniel Chang1,3, Yoonkyu Lee1,3, Connor M. Navis1, Alexandria L. Hillesheim1, Ian R. Nykaza1, Ngoc A. Ha1, Conner J. Hansen1, Fatma K. Karadag1, Rachel J. Bergerson4, Michael R. Verneris4, Matthew M. Meredith5, Matthew L. Schomaker5, Michael A. Linden6, Chad L. Myers7, David A. Largaespada3,4, and Zohar Sachs1,3 ABSTRACT ◥ Standard chemotherapy for acute myeloid leukemia (AML) onstrate that the transcriptional foundations of self-renewal and targets proliferative cells and efficiently induces complete remis- proliferation are distinct in LSCs as they often are in normal stem sion; however, many patients relapse and die of their disease. cells and suggest that therapeutic strategies that target self- Relapse is caused by leukemia stem cells (LSC), the cells with self- renewal, in addition to proliferation, are critical to prevent renewal capacity. Self-renewal and proliferation are separate relapse and improve survival in AML. functions in normal hematopoietic stem cells (HSC) in steady- stateconditions.Ifthesefunctionsarealsoseparatefunctionsin Significance: These findings define and functionally validate a LSCs, then antiproliferative therapies may fail to target self- self-renewal gene profile of leukemia stem cells at the single-cell renewal, allowing for relapse. We investigated whether prolifer- level and demonstrate that self-renewal and proliferation are ation and self-renewal are separate functions in LSCs as they distinct in AML. often are in HSCs. Distinct transcriptional profiles within LSCs of Graphical Abstract: http://cancerres.aacrjournals.org/content/ Mll-AF9/NRASG12V murine AML were identified using single-cell canres/80/3/458/F1.large.jpg. RNA sequencing. Single-cell qPCR revealed that these genes were also differentially expressed in primary human LSCs and normal Single-cell Self-renewal & Human single-cell LSC Primary AML Leukemia stem cells transcriptional profiling proliferation assays self-renewal genes human HSPCs. A smaller subset of these genes was upregulated in LSCs relative to HSPCs; this subset of genes constitutes “LSC- CD69 fi ” S100A4 speci c genes in human AML. To assess the differences between MYB ADA fi fi MRI1 these pro les, we identi ed cell surface markers, CD69 and CKS2 CD36, whose genes were differentially expressed between these profiles. In vivo mouse reconstitution assays resealed that only CD69High LSCs were capable of self-renewal and were poorly proliferative. In contrast, CD36High LSCs were unable to trans- Single-cell profiling reveals transcriptional mechanisms of self-renewal in AML. plant leukemia but were highly proliferative. These data dem- Introduction subset of cells, called leukemia stem cells (LSC), to recapitulate Acute myeloid leukemia (AML) is a lethal malignancy with a disease; these leukemia cells are endowed with long-term self- dismal overall survival rate. Standard AML chemotherapy induces renewal potential (2). Self-renewal is a process distinct from pro- complete remission in 60%–80% of patients; however, because of liferation in that it can give rise to pluripotent progenitors (stem high relapse rates, only 20%–30% of patients survive during the cells) in addition to daughter cells that differentiate into effector following two years (1). Relapse is caused by the ability of a small cells. Experimentally, self-renewal is defined as the ability to 1Division of Hematology, Oncology, and Transplantation, Department of Current address for R.S. LaRue, Minnesota Supercomputing Institute, University Medicine, University of Minnesota, Minneapolis, Minnesota. 2Next Generation of Minnesota, Minneapolis, Minnesota; current address for R.J. Bergerson, Analytics, Palo Alto, California. 3Masonic Cancer Center, University of Natural Sciences Department, Metropolitan State University, St. Paul, Minnesota; Minnesota, Minneapolis, Minnesota. 4Department of Pediatrics, University of and current address for M.R. Verneris, Department of Pediatrics, Medical Minnesota, Minneapolis, Minnesota. 5Molecular Lab, Department of Oncology and Hematology, University of Colorado School of Medicine, Center Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, for Cancer and Blood Disorders, Children's Hospital Colorado, University of Minnesota. 6Division of Hematopathology, Department of Laboratory Medicine Colorado Cancer Center, Aurora, Colorado. and Pathology, University of Minnesota, Minneapolis, Minnesota. 7Department of Corresponding Author: Zohar Sachs, University of Minnesota, 420 Delaware Computer Science and Engineering, University of Minnesota, Minneapolis, Street, SE, MMC 480, Minneapolis, MN 55455. Phone: 612-626-7055, Fax: 612- Minnesota. 624-6919; E-mail: [email protected] Note: Supplementary data for this article are available at Cancer Research Cancer Res 2020;80:458–70 Online (http://cancerres.aacrjournals.org/). doi: 10.1158/0008-5472.CAN-18-2932 K. Sachs, A.L. Sarver, K.E. Noble-Orcutt contributed equally to this article. Ó2019 American Association for Cancer Research. AACRJournals.org | 458 Downloaded from cancerres.aacrjournals.org on September 30, 2021. © 2020 American Association for Cancer Research. Published OnlineFirst November 29, 2019; DOI: 10.1158/0008-5472.CAN-18-2932 NRASG12V-Mediated Self-Renewal in Leukemia transplant leukemia to secondary recipients. Most leukemia cells expression profile at the single-cell level. Furthermore, these have high proliferative rates, but lack self-renewal capacity: they studies demonstrate that proliferation and self-renewal are cannot transplant leukemia to secondary recipients. In contrast, separate functions in immunophenotypically defined LSCs of this LSCs are a minor leukemia cell population that can give rise to the mouse model, as they are in HSCs, and suggest that curing AML full spectrum of leukemia subpopulations, including highly prolif- requires therapeutically targetingself-renewalinadditiontorapid erative daughter cells (2). Initial therapy for AML is efficient at proliferation. eliminating rapidly dividing leukemia cells (the bulk of the leukemia population), but self-renewing LSCs that survive can cause relapse. Our work seeks to identify molecular features of self-renewal so that Materials and Methods we can therapeutically target LSCs and prevent AML relapse. Experimental design A major challenge to studying LSCs is that the precise immu- Our study uses single-cell RNA sequencing to identify transcrip- fi nophenotypic pro le of true LSCs (leukemia cells that self-renew) tional heterogeneity within LSCs and define the self-renewing – varies by leukemia subtype and individual patients (3 6). subset within this compartment. See Experimental workflow figure fi The LSC-enriched compartment in AML has been identi ed (Supplementary Fig. S1A and S1B). We performed single-cell immunophenotypically by a cell surface protein expression RNA sequencing on the LSC-enriched compartment of our murine fi pro le. Thus far, many studies that query the molecular features leukemia model. As we previously demonstrated that self-renewal is of self-renewal have relied on cell surface proteins to identify an dependent on NRASG12V activity in this model, we also performed LSC-enriched population for evaluation. Relying on immunophe- single-cell RNA sequencing on the LSC-enriched compartment notypic markers of stemness may mask true features of self- after turning off NRASG12V transgene expression. To identify fi renewal. In immunophenotypically identi ed normal hematopoi- the functional contribution of each single-cell transcriptional pro- etic stem cells (HSC), oncogenic NRAS induced either prolifer- file, we performed in vivo assays of proliferation (using CellTrace ation or self-renewal, but these two functions were mutually labeling) and self-renewal (using in vivo leukemia reconstitution exclusive (7). We previously demonstrated that oncogenic NRAS assays). We also performed single-cell qPCR on primary human G12V fi (NRAS ) enforces a self-renewal gene expression pro le LSCs and normal bone marrow HSPCs to determine whether and is required for self-renewal of LSCs in a murine model of these cells express components of this single-cell self-renewal G12V G12V AML (Mll-AF9/NRAS ; refs. 8, 9). In this model, NRAS is a signature. tetracycline-repressible transgene under the control of the Vav1 promoter (10, 11): treatment with the tetracycline analogue, Murine single-cell RNA-sequencing analysis doxycycline, causes loss of NRASG12V expression and leads to Fragments per kilobase of exon model per million reads mapped disease remission. Moreover, in comparing the gene expression (FPKM) were modified by adding 0.1 to each value (to minimize the profiles of human and murine AML, we found that the effects of dividing by zero). Genes with FPKMs >0.1 in 50% of the NRASG12V-enforced self-renewal gene expression profile was cells were log transformed. The cells were hierarchically clustered expressed in multiple human AML self-renewal datasets (8). 2 using average linkage clustering using Pearson correlation. This report demonstrated that NRASG12V directs self-renewal in In Fig. 1A and B, the values were mean centered prior to clustering. LSCs as it does in normal HSCs (7);