The theory that cancer stem cells (CSCs), also referred to as tumour-initiating cells, constitute a small subpopu- lation of malignant tumour cells with enhanced capacities for self-renewal, metastatic dissemination and treatment resistance has been bolstered by an accumulating body of evidence. First described in acute myeloid leukaemia (AML), the postulated CSCs have since been isolated from other haematological malignancies and many types of solid cancers, and are hypothesized to form the clono- genic core of such tumours 1,2 . The cell of origin from which CSCs arise has not been unequivocally identified and almost certainly varies between different malignan- cies and perhaps even between individual tumours of the same histology. One hypothesis is that CSCs arise from more differentiated, non-stem cells that acquire stem cell-like properties after transformation, predominately owing to epithelial-to-mesenchymal transition (EMT) 3 ; another posits that CSCs originate from non-malignant stem cells through transformation induced by oncogenic somatic mutations 4 . Studies have also indicated that inflammation, and particularly inflammatory cytokines (such as interferons, tumour necrosis factor (TNF), IL-6 and IL-17), might have a role in inducing a CSC state 5 . Notably, multiple pools of CSCs are thought to be pres- ent within individual tumours, each of which can have unique biological characteristics 6 . Isolating and identifying CSCs remains a challenge; tumours contain heterogeneous cell populations, of which CSCs comprise only a small fraction (typically <1% in solid tumours), thus limiting our ability to detect them histologically 7 . Surrogate assays for CSCs include in vitro tumorsphere formation (a method commonly used in preclinical studies and that has been incorpo- rated into some clinical trials of CSC-targeting agents) and in vivo limiting-dilution tumorigenicity assays in immunocompromised mice, which is the gold standard method 8 . Importantly, particular cell-surface markers, including CD133, CD44 and epithelial cell-adhesion molecule, as well as aldehyde dehydrogenase enzy- matic activity are of value in the identification of CSCs 9 . However, discriminating between non-CSC tumour cells and true CSCs remains difficult as these markers are not Targeting signalling pathways and the immune microenvironment of cancer stem cells — a clinical update Joseph A. Clara 1,4 , Cecilia Monge 2,4 , Yingzi Yang 3 and Naoko Takebe 2 * Abstract | Cancer stem cells (CSCs) have important roles in tumour development, relapse and metastasis; the intrinsic self-renewal characteristics and tumorigenic properties of these cells provide them with unique capabilities to resist diverse forms of anticancer therapy, seed recurrent tumours, and disseminate to and colonize distant tissues. The findings of several studies indicate that CSCs originate from non-malignant stem or progenitor cells. Accordingly, inhibition of developmental signalling pathways that are crucial for stem and progenitor cell homeostasis and function, such as the Notch, WNT, Hedgehog and Hippo signalling cascades, continues to be pursued across multiple cancer types as a strategy for targeting the CSCs hypothesized to drive cancer progression — with some success in certain malignancies. In addition, with the renaissance of anticancer immunotherapy, a better understanding of the interplay between CSCs and the tumour immune microenvironment might be the key to unlocking a new era of oncological treatments associated with a reduced propensity for the development of resistance and with enhanced antimetastatic activity, thus ultimately resulting in improved patient outcomes. Herein, we provide an update on the progress to date in the clinical development of therapeutics targeting the Notch, WNT, Hedgehog and Hippo pathways. We also discuss the interactions between CSCs and the immune system, including the potential immunological effects of agents targeting CSC-associated developmental signalling pathways, and provide an overview of the emerging approaches to CSC-targeted immunotherapy. 1 National Heart Lung and Blood Institute, NIH, Bethesda, MD, USA. 2 Division of Cancer Treatment and Diagnosis, National Cancer Institute, NIH, Bethesda, MD, USA. 3 Department of Developmental Biology, Harvard School of Dental Medicine, Dana-Farber/ Harvard Cancer Center, Boston, MA, USA. 4 These authors contributed equally: Joseph A. Clara, Cecilia Monge *e-mail: takeben@ mail.nih.gov https://doi.org/10.1038/ s41571-019-0293-2 REVIEWS www.nature.com/nrclinonc 204 | APRIL 2020 | VOLUME 17