Helen M. Blau

The Evolving Concept of a Stem Cell: Entity or Function?

We wish to thank A. Banfi, B. Colyear, and N. Gewertz for expert assistance with artwork. We are indebted to M. Baron, J. Botas, G. Cossu, E. Gussoni, Y.N. Jan, T. Palmer, T. Tlsty, B. Wold, and members of our laboratory for in-depth critique of the manuscript. We gratefully acknowledge review of the tissue-specific material by E. Fuchs, M. Kay, P. Khavari, A. Oro, and L. Reid. We apologize in advance to those whose work we were not able to cover due to space constraints. This work was supported by a Lutheran Fellowship and an NIH predoctoral training grant GM07149 to T.R.B. and by NIH grants CA59717, AG09521, HD 18179, and HL65572 to H.M.B.

Designing materials to direct stem-cell fate

Proper tissue function and regeneration rely on robust spatial and temporal control of biophysical and biochemical microenvironmental cues through mechanisms that remain poorly understood. Biomaterials are rapidly being developed to display and deliver stem-cell-regulatory signals in a precise and near-physiological fashion, and serve as powerful artificial microenvironments in which to study and instruct stem-cell fate both in culture and in vivo. Further synergism of cell biological and biomaterials technologies promises to have a profound impact on stem-cell biology and provide insights that will advance stem-cell-based clinical approaches to tissue regeneration.

Substrate Elasticity Regulates Skeletal Muscle Stem Cell Self-Renewal in Culture

Environment Matters Stem cells isolated from muscle can be used for muscle regeneration, but only if the stem cells are fresh. Under standard cell culture conditions in the laboratory, muscle stem cells fail to proliferate efficiently and lose their regenerative capacity. Gilbert et al. (p. 1078, published online 15 July; see the Perspective by Bhatia) built an in vitro–culture system that resembles the physical characteristics in which muscle stem cells normally reside: a squishy elastic bed (rather than the hard slab of a plastic culture flask). Laminin tethered to hydrogels was used to generate substrates of varying elasticity. When cultured on these substrates, muscle stem cells remained undifferentiated and were able to support muscle regeneration when transplanted back into mice. Muscle stem cells prefer a soft substrate. Stem cells that naturally reside in adult tissues, such as muscle stem cells (MuSCs), exhibit robust regenerative capacity in vivo that is rapidly lost in culture. Using a bioengineered substrate to recapitulate key biophysical and biochemical niche features in conjunction with a highly automated single-cell tracking algorithm, we show that substrate elasticity is a potent regulator of MuSC fate in culture. Unlike MuSCs on rigid plastic dishes (~106 kilopascals), MuSCs cultured on soft hydrogel substrates that mimic the elasticity of muscle (12 kilopascals) self-renew in vitro and contribute extensively to muscle regeneration when subsequently transplanted into mice and assayed histologically and quantitatively by noninvasive bioluminescence imaging. Our studies provide novel evidence that by recapitulating physiological tissue rigidity, propagation of adult muscle stem cells is possible, enabling future cell-based therapies for muscle-wasting diseases.

Biological Progression from Adult Bone Marrow to Mononucleate Muscle Stem Cell to Multinucleate Muscle Fiber in Response to Injury

Adult bone marrow-derived cells (BMDC) are shown to contribute to muscle tissue in a step-wise biological progression. Following irradiation-induced damage, transplanted GFP-labeled BMDC become satellite cells: membrane-ensheathed mononucleate muscle stem cells. Following a subsequent exercise-induced damage, GFP-labeled multinucleate myofibers are detected. Isolated GFP-labeled satellite cells are heritably myogenic. They express three characteristic muscle markers, are karyotypically diploid, and form clones that can fuse into multinucleate cells in culture or into myofibers after injection into mouse muscles. These results suggest that two temporally distinct injury-related signals first induce BMDC to occupy the muscle stem cell niche and then to help regenerate mature muscle fibers. The stress-induced progression of BMDC to muscle satellite cell to muscle fiber results in a contribution to as many as 3.5% of muscle fibers and is due to developmental plasticity in response to environmental cues.

VEGF Gene Delivery to Myocardium Deleterious Effects of Unregulated Expression

BACKGROUND Vascular endothelial growth factor (VEGF) is being investigated for therapeutic angiogenesis in ischemic myocardium. Primarily, transient delivery systems have been tested. The goal of this study was to investigate the effects of continuous expression of VEGF in myocardium by use of myoblast-mediated delivery. METHODS AND RESULTS Primary murine myoblasts (5 x 10(5) cells in 10 microL of PBS with 0.5% BSA) expressing both the murine VEGF gene and the beta-galactosidase (beta-gal) gene from a retroviral promoter were implanted in the ventricular wall of immunodeficient mice (n=11) via a subdiaphragmatic approach. Control immunodeficient mice (n=12) were injected with the same number of myoblasts expressing only the beta-gal gene. Between days 14 and 16, surviving mice were euthanized and the hearts processed for histology. In the experimental group, 11 of 11 mice demonstrated failure to thrive by day 13; 5 deaths occurred between days 8 and 15. There were no complications in the control mice. Histochemistry documented successful implantation of myoblasts (positive beta-gal reaction product) in 6 of 6 surviving experimental mice and 12 of 12 controls. Histology disclosed intramural vascular tumors resembling hemangiomas in the VEGF-myoblast-injected myocardium in 6 of 6 surviving mice. beta-Gal-expressing cells were present at the site of the vascular tumors. Immunohistochemistry localized abundant endothelial nitric oxide synthase and CD31 (platelet and endothelial cell adhesion molecule) within the lesion, consistent with the presence of endothelial cells. CONCLUSIONS In this model, unregulated continuous expression of VEGF is associated with (1) a high rate of failure to thrive/death and (2) formation of endothelial cell-derived intramural vascular tumors in the implantation site. These results underscore the importance of regulating VEGF expression for therapeutic angiogenesis.

Dermatologist-level classification of skin cancer with deep neural networks

Skin cancer, the most common human malignancy, is primarily diagnosed visually, beginning with an initial clinical screening and followed potentially by dermoscopic analysis, a biopsy and histopathological examination. Automated classification of skin lesions using images is a challenging task owing to the fine-grained variability in the appearance of skin lesions. Deep convolutional neural networks (CNNs) show potential for general and highly variable tasks across many fine-grained object categories. Here we demonstrate classification of skin lesions using a single CNN, trained end-to-end from images directly, using only pixels and disease labels as inputs. We train a CNN using a dataset of 129,450 clinical images—two orders of magnitude larger than previous datasets—consisting of 2,032 different diseases. We test its performance against 21 board-certified dermatologists on biopsy-proven clinical images with two critical binary classification use cases: keratinocyte carcinomas versus benign seborrheic keratoses; and malignant melanomas versus benign nevi. The first case represents the identification of the most common cancers, the second represents the identification of the deadliest skin cancer. The CNN achieves performance on par with all tested experts across both tasks, demonstrating an artificial intelligence capable of classifying skin cancer with a level of competence comparable to dermatologists. Outfitted with deep neural networks, mobile devices can potentially extend the reach of dermatologists outside of the clinic. It is projected that 6.3 billion smartphone subscriptions will exist by the year 2021 (ref. 13) and can therefore potentially provide low-cost universal access to vital diagnostic care.

Cytoplasmic activation of human nuclear genes in stable heterocaryons

We have induced the stable expression of muscle-specific genes in human nonmuscle cells. Normal diploid human amniocytes were fused with differentiated mouse muscle cells by using polyethylene glycol. The fusion product, a stable heterocaryon in which the parental cell nuclei remained distinct, did not undergo division and retained a full complement of chromosomes. This is in contrast with typical interspecific hybrids (syncaryons), in which the parental nuclei are combined and chromosomes are progressively lost during cell division. The human muscle proteins, myosin light chains 1 and 2, MB and MM creatine kinase and a functional mouse-human hybrid MM enzyme molecule were detected in the heterocaryons. Synthesis of these proteins was evident 24 hr after fusion and increased in a time-dependent manner thereafter. Our results indicate that differentiated mouse muscle nuclei can activate human muscle genes in the nuclei of a cell type in which they are not normally expressed, and that this activation occurs via the cytoplasm. The activators are still present in cells which have already initiated differentiation, are recognized by nuclei of another species, and do not diffuse between unfused cells. The reprogrammed amniocyte nuclei of stable heterocaryons provide a unique system in which to study the mechanisms regulating gene expression during cell specialization.

Reprogramming towards pluripotency requires AID-dependent DNA demethylation

Reprogramming of somatic cell nuclei to yield induced pluripotent stem (iPS) cells makes possible derivation of patient-specific stem cells for regenerative medicine. However, iPS cell generation is asynchronous and slow (2–3 weeks), the frequency is low (<0.1%), and DNA demethylation constitutes a bottleneck. To determine regulatory mechanisms involved in reprogramming, we generated interspecies heterokaryons (fused mouse embryonic stem (ES) cells and human fibroblasts) that induce reprogramming synchronously, frequently and fast. Here we show that reprogramming towards pluripotency in single heterokaryons is initiated without cell division or DNA replication, rapidly (1 day) and efficiently (70%). Short interfering RNA (siRNA)-mediated knockdown showed that activation-induced cytidine deaminase (AID, also known as AICDA) is required for promoter demethylation and induction of OCT4 (also known as POU5F1) and NANOG gene expression. AID protein bound silent methylated OCT4 and NANOG promoters in fibroblasts, but not active demethylated promoters in ES cells. These data provide new evidence that mammalian AID is required for active DNA demethylation and initiation of nuclear reprogramming towards pluripotency in human somatic cells.

Argonaute 2/RISC resides in sites of mammalian mRNA decay known as cytoplasmic bodies

RNA interference (RNAi) is an important means of eliminating mRNAs, but the intracellular location of RNA-induced silencing complex (RISC) remains unknown. We show here that Argonaute 2, a key component of RISC, is not randomly distributed but concentrates in mRNA decay centres that are known as cytoplasmic bodies. The localization of Argonaute 2 in decay centres is not altered by the presence or absence of small interfering RNAs or their targeted mRNAs. However, RNA is required for the integrity of cytoplasmic bodies because RNase eliminates Argonaute 2 localization. In addition, Argonaute 1, another Argonaute family member, is concentrated in cytoplasmic bodies. These results provide new insight into the mechanism of RNAi function.

Self-renewal and expansion of single transplanted muscle stem cells

Adult muscle satellite cells have a principal role in postnatal skeletal muscle growth and regeneration. Satellite cells reside as quiescent cells underneath the basal lamina that surrounds muscle fibres and respond to damage by giving rise to transient amplifying cells (progenitors) and myoblasts that fuse with myofibres. Recent experiments showed that, in contrast to cultured myoblasts, satellite cells freshly isolated or satellite cells derived from the transplantation of one intact myofibre contribute robustly to muscle repair. However, because satellite cells are known to be heterogeneous, clonal analysis is required to demonstrate stem cell function. Here we show that when a single luciferase-expressing muscle stem cell is transplanted into the muscle of mice it is capable of extensive proliferation, contributes to muscle fibres, and Pax7+luciferase+ mononucleated cells can be readily re-isolated, providing evidence of muscle stem cell self-renewal. In addition, we show using in vivo bioluminescence imaging that the dynamics of muscle stem cell behaviour during muscle repair can be followed in a manner not possible using traditional retrospective histological analyses. By imaging luciferase activity, real-time quantitative and kinetic analyses show that donor-derived muscle stem cells proliferate and engraft rapidly after injection until homeostasis is reached. On injury, donor-derived mononucleated cells generate massive waves of cell proliferation. Together, these results show that the progeny of a single luciferase-expressing muscle stem cell can both self-renew and differentiate after transplantation in mice, providing new evidence at the clonal level that self-renewal is an autonomous property of a single adult muscle stem cell.