A Stem Cell Perspective on Cellular Engineering

See allHide authors and affiliations

Science  08 Nov 2013:
Vol. 342, Issue 6159, pp. 700-702
DOI: 10.1126/science.1238363

A fundamental enigma in modern biology concerns the molecular rules that govern how cells establish and maintain identity during development. These rules are the key to generating therapeutic cell types in vitro and the foundation of regenerative medicine. The isolation of embryonic stem cells (ESCs) has enabled scientists to recapitulate the process of embryonic development in a dish, by directing differentiation of ESCs with combinations of morphogens and growth factors to mimic embryonic development. These experiments assume that cell and tissue fates evolve along specific paths, and once established, remain fixed. But the advent of cellular reprogramming has fundamentally altered our view of the stability of cell identity, and dramatic demonstrations of the interconversions of mature cell types have introduced the provocative idea that cell identity can be engineered to play beneficial therapeutic roles.

The discovery of induced pluripotency taught us that a somatic cell can be reprogrammed to a pluripotent state (an induced pluripotent stem cell, iPSC) by enforced expression of several transcription factors (1). Stable cell identity depends on the existence of preferred “attractor states” of the epigenome, which represent stable “valleys” within an otherwise dynamic landscape of cellular phenotypes. Cell identities often hinge on bistable switches, which are governed by sets of transcription factors that promote and reinforce one lineage program while antagonizing alternatives, as in the segregation of the inner cell mass and trophectoderm in the early embryo. Development proceeds until cell identity is stabilized in a terminally differentiated state. A challenge to the notion of fixed cell identities is the recent finding that the expression of heterologous lineage factors can enable direct conversions between cell types (2). Generally, direct conversion occurs without involving normal developmental intermediates. Although direct conversion aims to reproduce physiologic mechanisms of lineage specification, it can lead to aberrant cell types without a clear equivalent in nature. Novel computational methods will be needed to assess the fidelity of reported conversions and to diagnose which gene sets remain incompletely converted or aberrantly activated. Should these engineered cells function effectively, they may prove highly valuable for medical applications, but their safety must first be proven.

Engineering cell identity.

Directed differentiation of ESCs and iPSCs relies on morphogens and growth factors to mimic embryonic development. Direct conversion uses transcription factors to force somatic cells to transit between cell states, generally without involving normal developmental intermediates. Differentiation or conversion can follow one of two paradigms. Functionally mature cell types can be produced in a dish, but they may have a limited life span after transplantation in vivo. Alternatively, somatic stem cells for the target tissue are derived and transplanted; they are then stably maintained in their native niches, self-renew indefinitely, and differentiate to tissue-specific cell types.


Directed differentiation from pluripotent stem cells (PSCs) and direct conversion from other somatic cells have emerged as two powerful paradigms for manipulating cell fate (see the figure) (2, 3). What has received less attention is the place that somatic stem cells occupy in these paradigms. Most mature cell types in the body are not self-renewing; they are either continuously replenished from somatic stem cells (e.g., hematopoietic, intestinal, and epidermal) or are replaced slowly or intermittently in response to injury (e.g., muscle satellite cells). There are important differences that concern the rate of turnover (high turnover in blood, skin, and intestine, versus low turnover in the heart and brain), and exceptions (some cell types, such as pancreatic beta cells and memory T cells, renew without stem cells). For many tissues, an important question is whether, for therapeutic purposes, the goal of directed differentiation or direct conversion should be to generate functionally mature cells or to generate immature somatic stem cells.

Directed differentiation or direct conversion are typically aimed at generating a mature cell type, such as a contracting cardiomyocyte or a dopaminergic neuron. Although this approach is most commonly pursued today, its fundamental limitation is that mature cells may be postmitotic or lack significant proliferative potential, which limits both the number of cells that can be obtained in a dish and their ability to expand or even simply persist in situ after transplantation. Cardiomyocytes and neurons are long-lived cell types, which favors their maintenance in vivo, yet efforts to deliver these cells into the heart muscle or brain are hampered by the inability of mature cells to properly integrate and persist in the tissue.

Another approach is to generate somatic stem cells that can be expanded and matured in a dish or transplanted into the host tissue. Target stem cells include hematopoietic (HSCs), intestinal (ISCs), mesenchymal (MSCs), neural, cardiac progenitor, skeletal muscle progenitors and satellite cells, and bronchio-alveolar lung stem cells. Instead of promoting maturation, protocols strive to capture and propagate the stem cell state. The success of transplantation is facilitated by the natural capacity of stem cells for long-term persistence and renewal in their native niches, ensuring a lasting benefit of the graft. There is also increasing evidence for many cell types that in vivo maturation produces cells with improved function relative to those matured in a dish (4). This strategy is promising, yet it faces a number of difficulties. Stem cells are rare, and mechanisms of self-renewal are poorly understood, complicating efforts to expand and maintain stem cells.

These challenges are especially evident in attempts to regenerate the hematopoietic lineage from PSCs. Because of their ability to fully reconstitute the blood system upon transplantation, HSCs are an extremely valuable therapeutic cell type. Numerous differentiation protocols have been established that attempt to mimic the conditions of hematopoietic ontogeny (1, 5, 6). Direct conversion from fibroblasts using a combination of four transcription factors—cFos, Gata2, Gfi1b, and Etv6—has also recently been reported (7). These protocols generate large numbers of hematopoietic progenitors, but not bona fide HSCs. Even if HSCs could be generated, no conditions exist to faithfully preserve stem cell potential during prolonged culture. The long-standing struggle to generate transplantable HSCs highlights the need for novel approaches to generating somatic stem cells.

One alternative may be conversion from related lineages that minimize the “epigenetic distance” to a desired cell type, providing the context for more precise cell fate alterations. Myeloid precursors differentiated from human PSCs can be respecified into transplantable multipotential progenitors with the combination of the transcription factors ERG, HOXA9, and RORA (8). A similar approach may be applied to other tissues by introducing stem cell transcription factors into progenitors or transient amplifying cells from PSCs or primary sources. In vitro culture conditions are impoverished relative to the three-dimensional, dynamic milieu of the developing embryo. This environment may have to be recapitulated in vitro to capture and maintain stem cells. For instance, ISCs require Wnt signals from neighboring Paneth cells; ISCs aggregated with Paneth cells efficiently form three-dimensional organoids (9). Hepatic endoderm differentiated from human PSCs can aggregate with primary MSCs and human umbilical vein endothelial cells, recapitulating organogenesis in a dish (10).

In cases where the niche is not known or is difficult to recapitulate in vitro, conversion may be carried out directly in vivo. Delivery of the set of transcription factors Gata4, Mef2c, and Tbx5 into the mouse heart induces the conversion of cardiac fibroblasts into cardiomyocytes that show improved function relative to in vitro reprogrammed counterparts (4). However, translation to human may not always be direct. For instance, Hoxb4 converts blood precursors from mouse ESCs into HSC-like cells (11), yet extensive efforts to adapt HOXB4 for human cells have been largely unsuccessful (12). The advent of powerful genome-editing technologies enables the creation of transgenic human lines harboring defined factors or stem cell reporters. Combined with improved xenotransplantation models, engineering directly in human cells with functional validation in engrafted mice is an attractive approach.

Moving forward, the stem cell research community must creatively apply directed differentiation and direct conversion toward engineering clinically valuable cells, targeting the generation of either mature functional cells or stem cells, depending on the anticipated clinical application and guided by the cell type and tissue of interest. For long-lived cells such as cardiomyocytes and neurons, integration of mature cells into the tissue remains a viable option. Still, the stem cell approach should be explored, because functional tissue integration may be more permissive for neural or cardiac progenitors. For short-lived tissues such as blood, mesenchyme, skin, or intestinal epithelium, the generation of somatic stem cells will be a prerequisite for stable engraftment and prolonged tissue reconstitution. Such advances will require novel markers and reporter lines, deeper understanding of stem cell-specific transcription factors, and screening strategies formulated to derive and detect rare stem cells. The fundamental demonstration of the past decade that cell identity can be molded to our specifications has created an unprecedented opportunity to create rare patient-specific cell types and even tissues. The decade to come will establish whether this revolution in basic science will have a lasting impact on medicine.


View Abstract

Stay Connected to Science

Navigate This Article