Introduction

In this chapter we outline the concepts and practical considerations in understanding and developing defined culture systems for the controlled in vitro differentiation of mouse embryonic stem (mES) cells.

When looking at the history of mES-cell in vitro differentiation, one is struck by the level of alchemy that was introduced, and at times still is used, to obtain desired cell types. However, a fundamental difference between alchemy and science is that science is based on understanding; in designing ES-cell differentiation experiments it is essential that we have a total grasp of the question, and use reagents that are defined and reliable. In examining the differentiation of ES cells into specific, differentiated cell types we must define the system by using reagents that allow interpretation of the results and, equally important, ensure reproducibility. Ideas can thus be proposed and tested, data interpreted, and the information used by others to extend the work. It is only by progressing from alchemy to science that the full potential of ES cells can be exploited and developed coherently.

The emergence of in vitro cell culture techniques occurred in the early 20th century (1, 2) and rapidly lead to the establishment of the first human tumour cell lines, for example HeLa (3). These achievements reflected considerable advances in understanding the biophysical and biochemical parameters for cell survival and maintenance. In working with cells in vitro it is evident that the key to success is the immediate microenvironment; that is, cell culture systems must provide nutrients, survival factors, protection from negative effects of the atmosphere, and dilution of waste products. However, eukaryotic cells and their culture still represent an intricate biological system of a subtle complexity far beyond our current understanding. Therefore we must even now recognize that any analytical cell culture-based approach is still a balance between what we understand and what remains to be discovered.

The history of in vitro cell differentiation began during the 1970s, when embryonal carcinoma (EC) cell lines were first isolated. EC cells show an amazing ability to differentiate in vitro, and can perhaps contribute to the formation of multiple cell types in chimeras. However, in general EC cells exhibit limited, or possibly predetermined, patterns of differentiation in vitro. In the early 1980s mouse ES (mES) cells were isolated (see Chapter 1). These cells are regarded as pluripotent, showing a far greater capacity to differentiate in vitro and in vivo than EC cells. Linked with this greater differentiation capacity of mES cells comes more stringent cell culture requirements, to sustain their proliferation as undifferentiated and pluripotent cells (4). The maintenance and differentiation of ES cells from diverse species currently presents some of the greatest challenges to researchers studying stem cells and their controlled differentiation.

Stem cells, by definition, renew and/or differentiate. The first mES cell lines were isolated and were maintained in an undifferentiated state by coculture with feeders cells and/or in the presence of cell-conditioned media (5). With the discovery that leukaemia inhibitory factor (LIF) prevents mES-cell differentiation in culture (at least for mES cells derived from the 129 mouse substrains), mES cell culture became simpler (6). Under conventional culture conditions using medium supplemented with batch-selected fetal calf serum (FCS) and LIF or similar defined factors (e.g. CNTF, Il-11 or OSM (7-10), and see Chapter 2), mES cells are maintained poised between self-renewal and differentiation, with self-renewal being favoured. In the absence of LIF and in the presence of FCS, differentiation dominates over self-renewal. It is the ability specifically to control the nature and outcome of this differentiation process, producing any desired cell type upon command, that is the Holy Grail for many stem-cell researchers.

In vitro differentiation of mES cells is performed most frequently via their aggregation, forming clumps of cells known as 'embryoid bodies' (EBs). These structures can contain experimental sources of intermediate progenitor cells for the three germ layers. In EBs, mES cells undergo transitions to ectoderm, meso-derm, and to visceral and parietal endoderm, mimicking to varying degrees events at gastrulation (from day 4.5 to 7.5 p.c. in the mouse). However, the factors and events occurring within these complex structures are poorly defined and not well understood. Understanding is further complicated as these structures are highly dynamic in nature, with cellular differentiation, survival, and differential expansion of differentiated derivatives all being sensitive to numerous parameters, most of which are not known. Our approach to unravelling these complexities has been to refine in vitro culture systems, allowing controlled manipulation of germ-layer differentiation from mES cells (11-15).

Independently during the mid 1980s, Tom Doetschman and Anna Wobus studied in vitro differentiation of mES cells. Both showed that although EB-mediated induction of mES-cell differentiation appeared to be simple, the efficiency of differentiation and the nature of cell types obtained were difficult to regulate or predict (16, 17). The lack of repeatable, controlled in vitro differentiation severely limited the use of mES cells as a tool in understanding embryonic development; it also limited their exploitation in pharmaceutical testing and the burgeoning field of cell therapy.

More recently, we demonstrated that mES cells could be routinely and efficiently differentiated into mesoderm and various haematopoietic lineages in media containing 10-15% batch-selected FCS (11, 12). Additionally, in the case of mesoderm and subsequent haematopoiesis, the sequence of events mimics aspects of embryonic development in vivo (8). The critical, simple medium additive that made this approach reproducible was monothioglycerol (MTG), or b-mercaptoethanol also was shown to be effective; that is, an active reducing agent was shown to be crucial to the survival of differentiating cells. These findings indicate that free-radical scavenging is an essential environmental need for reproducible mES-cell in vitro differentiation. These data also revealed that a principal 'driver' of mES-cell differentiation in vitro is serum (i.e. batch-selected FCS), present in the medium at 10-15%. It was consequently shown that serum effects are generally overriding, so that the addition of many growth factors appears to have little or no influence on the initial development of either mesoderm or subsequent haematopoietic progenitors. However once progenitors develop, exogenous factors do influence their survival and also the subsequent expansion of differentiated cell types, such as erythrocytes, macrophages, and mast cells, all of which can be specifically enhanced (12, 18).

The finding that FCS effects are so dominant in mES cells during in vitro differentiation was the stimulus to devise a cell culture system where all components of the medium are defined and described completely; that is, a chemically defined medium (CDM) (13-15) in which FCS and (later) bovine serum albumin (BSA) were replaced by chemically defined reagents.

Differentiation experiments using CDM allowed, for the first time, separation of serum effects from those of specific, exogenously added growth factors. Using this approach we conclusively demonstrated that mesoderm is induced by the TGF-b superfamily members, bone morphogenetic protein-2, -4, -7 (BMP-2, BMP-4, BMP-7), or activin A (Figure 1). Furthermore, we found that these factors initiated a process that resembles primitive-streak formation, followed rapidly by haematopoietic cell formation and expansion (13-15). Using CDM, important precedents were established: (i) mES cells, whilst cultured as EBs in basal CDM, are responsive to induction of differentiation by treatment with exogenous growth factors; and (ii) in this system, mES cells can be directed towards neuroec-todermal or mesodermal pathways (13-15).

A constructive lesson on the masking effect of serum (or plasma-derived serum) can be drawn from the studies of EB-mediated haematopoiesis. Using FCS to supplement media, differentiation can be driven towards mesoderm and, depending upon the serum batch, erythroid-myeloid pathways. Evidence from non-mammalian systems (e.g. Xenopus and Zebrafish) had indicated that BMP is crucial in mesoderm induction; however, in FCS-based mES-cell differentiation systems, treatment with this factor appeared to have no observable effect. By using serum-free, defined medium it was possible to demonstrate unequivocally

Differentiation days

No factors

Activin A 2 ng/ml

Differentiation days

No factors

Activin A 2 ng/ml

Figure 1 Timecourse analysis of expression of a mesoderm-specific gene in mES cells developing as EBs in CDM plus BMP-2, BMP-4, or activin A (13-15, 29). Line CCE mES cells were seeded for EB formation in basal CDM, or in CDM plus0.25 ng/ml BMP-4, 2 ng/ml BMP-4, or 2 ng/ml activin A. Semiquantitative RT-PCR for T(Brachyury) expression was used to monitor mesoderm development. In this example the approximate amounts of cDNA were adjusted using a housekeeping gene, hypoxanthine guanine phosphoribosyl transferase (HPRT). PCR products were separated electrophoretically, Southern blotted, and hybridized using specific probes. 'Day 0' used undifferentiated mES cells. The negative controls were water, and the positive control cDNA from EBs cultured for 4d in FCS-supplemented medium (27). In EBs cultured with 2 ng/ml activin A, expression of Tbecame detectable by 2d of differentiation, was significantly elevated by 5d, and continued to be expressed over the period examined. With 2 ng/ml BMP-4, Texpression was strongly induced at 3-4d and then diminished; this is reminiscent of the pulse of mesodermal gene expression occurring during primitive-streak formation in vivo. Interestingly, at the lower concentration of BMP-4, 0.25 ng/ml, induction of Texpression occurred later, at 4d, and took longer to diminish.

Brachyury HPRT

Brachyury HPRT

Brachyury HPRT

Brachyury HPRT

Figure 1 Timecourse analysis of expression of a mesoderm-specific gene in mES cells developing as EBs in CDM plus BMP-2, BMP-4, or activin A (13-15, 29). Line CCE mES cells were seeded for EB formation in basal CDM, or in CDM plus0.25 ng/ml BMP-4, 2 ng/ml BMP-4, or 2 ng/ml activin A. Semiquantitative RT-PCR for T(Brachyury) expression was used to monitor mesoderm development. In this example the approximate amounts of cDNA were adjusted using a housekeeping gene, hypoxanthine guanine phosphoribosyl transferase (HPRT). PCR products were separated electrophoretically, Southern blotted, and hybridized using specific probes. 'Day 0' used undifferentiated mES cells. The negative controls were water, and the positive control cDNA from EBs cultured for 4d in FCS-supplemented medium (27). In EBs cultured with 2 ng/ml activin A, expression of Tbecame detectable by 2d of differentiation, was significantly elevated by 5d, and continued to be expressed over the period examined. With 2 ng/ml BMP-4, Texpression was strongly induced at 3-4d and then diminished; this is reminiscent of the pulse of mesodermal gene expression occurring during primitive-streak formation in vivo. Interestingly, at the lower concentration of BMP-4, 0.25 ng/ml, induction of Texpression occurred later, at 4d, and took longer to diminish.

a critical role for BMP-4 or BMP-2 in mesoderm, and subsequent erythropoietic, differentiation. Thus, it is the 'inherent' factors in FCS that can induce these differentiation processes as a matter of course under conventional in vitro culture conditions (15).

Nakayama et al. extended these results, confirming that under serum-free conditions BMP-4 is essential for erythroid-myeloid differentiation; and also found that the FLT1 ligand, VEGF, acts synergistically with BMP-4, enhancing BMP-4-dependent lympho-haematopoietic cell production (19). This group published that BMP-4 causes differentiation of mES cells, cultured as EBs under serum-free conditions, into non-committed mesodermal precursor cells (possibly the presumptive haemangioblast); and that subsequent differentiation into lympho-haematopoietic progenitor cells requires the sequential activity of VEGF. This contradicted other reports, where FCS-supplemented medium was used and where an inhibitory effect of VEGF on erythropoietic differentiation had been reported.

In summary, although developmentally important genes are being identified in increasing numbers, the actual mechanisms of cellular differentiation, the precise effects of growth factor signalling molecules, and the nature of sequential progenitor cells, have yet to be fully elucidated. Defined culture systems offer an experimental framework in which these highly complex processes can begin to be successfully dissected.

0 0

Post a comment