Innovation in the Culture and Derivation of Pluripotent Human Stem Cells

Feeder Cells

Originally, primary mouse embryonic fibroblasts (MEFs), and later immortalized MEFs (STO line), were used for the derivation and propagation of hESCs in an undifferentiated state [1,2]. Due to concerns about hESC contact with xenogeneic cells, a number of human cell types have been subsequently examined as alternative feeder cells for hESC culture. Fibroblasts from different stages of development (fetal, neonatal, adult) and tissue sources (skin, muscle, placenta, uterus) are reportedly capable of supporting hESC undifferentiated growth, although certain cells support hESC maintenance better than others (for a review of feeder cells for hESC culture see [3]). In addition to allogeneic cell sources, fibroblast-like cells that spontaneously differentiate in hESC cultures can also be isolated and propagated to serve as an autogeneic source of feeder cells [4]. The differences in the performance of various feeder cells to support the pluripotency and self-renewal of hESCs can be attributed to the production of different types and amounts of secreted molecules. For example, MEFs typically produce more activin A and less basic fibroblast growth factor (bFGF) than human foreskin fibroblasts, which produce variable amounts of bFGF among different lines [5].

Culture Media Composition

Soon after the initial derivation of hESC lines, the first feeder-free culture system was described. In lieu of feeder layers, hESCs were grown for extended passages on Matrigel-coated substrates in the presence of MEF-conditioned medium [6]. Recent interrogation of the complex compositions of media conditioned by fibroblasts has provided insight into specific factors capable of supporting hESC pluripotency in vitro and facilitated the transition from feeder cells to defined media. By performing comparative proteomic analysis on media before and after conditioning by human fetal, neonatal and mouse embryonic fibroblasts, 175 proteins were uniquely identified, of which 34 were shared among the 3 cell types [7]. Many of the proteins identified, such as Activin/Inhibin, insulin-like growth factor-1 (IGF1) and BMP/TGF-beta1, were independently identified to play an active role in pluripotent hESC growth. Molecules secreted into the media by primary MEFs and a MEF line incapable of supporting hESCs in an undifferentiated state were also compared to identify pluripotent maintenance factors. As a result, a mixture of 6 recombinant proteins capable of supporting hESC growth (pigment epithelium-derived factor, plasminogen activator inhibitor, insulin-like growth factor binding proteins 2 and 7, monocyte chemoattractant protein 1, and interleukin 6) could be substituted for the more complex MEF conditioned media when culturing cells on fibronectin [8].

Based in part on analysis of conditioned media, a number of individual soluble factors, as well as combinations of molecules, have been identified to be important regulators of pluripotent stem cell self-renewal (Figure 1). Traditionally, bFGF has been the primary supplement added to hESC culture media according to original derivation and culture protocols for undifferentiated cell growth. Several years ago, several independent groups reported that bFGF was necessary and sufficient for hESC growth under feeder-free conditions, even without MEF-conditioned media [9–12]. By increasing the concentration of bFGF to 100 ng/ml (~6 nM, nearly 25x the amount added to MEF CM), hESCs could be sustained in a pluripotent state similar to CM [13]. The requirement of bFGF concentrations substantially above the K_d (< 1 nM) may be an effect of the low thermal stability of bFGF in culture conditions or alternatively may permit signaling through lower-affinity interactions [13].

In addition to bFGF, TGFβ/Activin/Nodal signaling via the SMAD2/3 pathway has also been demonstrated to play a significant, if not critical role in regulating the pluripotency of hESCs in the absence of feeders layers [14–16]. The combination of Activin with bFGF can sufficiently maintain hESC self-renewal even in the absence of feeder cells, CM, and serum replacement [17], and Activin A treatment induces expression of numerous pluripotent genes, such as Oct4 and Nanog, and stimulatory growth factors, such as Nodal, Wnt3, bFGF and FGF8 [18]. Inhibiting the bioactivity of potent morphogens, such as bone morphogenic proteins, by a specific antagonist, like Noggin, may also enhance the maintenance of hESCs in a pluripotent state [12,19]. More recently, analysis of specific growth factor receptor expression and signal transduction implicated the IGF signaling pathway as a direct mediator of hESC growth, whereas bFGF appeared to play an indirect role [20,21].

Although most hESC self-renewal studies have focused on the effects of growth factor proteins, recent evidence indicates that lipid molecules, abundant in serum and knockout serum replacement, are also important factors capable of maintaining hESC pluripotency. The addition of sphingosine-1-phosphate (S1P), in combination with platelet derived growth factor (PDGF), to serum-free media inhibited the spontaneous differentiation of hESCs grown on MEFs [22]. Likewise, supplementation of hESC culture media with sphingolipid metabolites, such as S1P or ceramide, appear to benefit the maintenance of hESC cultures by enhancing proliferation, as well as preventing hESC apoptosis and perhaps selectively promoting apoptosis of differentiating cells [23–25]. Due to their hydrophobic nature, introduction of lipids to aqueous media requires a hydrophilic carrier molecule, often albumin, in order to be remain soluble. A recent comparison of hESC cultures treated with either lipid-rich or lipid-poor fractions of bovine serum albumin (BSA) clearly demonstrated that only the lipid-rich BSA stimulated self-renewal of karyotypically normal hESCs in a pluripotent state [26].

Toward Defined, Humanized Culture Conditions

The use of hESCs in clinical trials and therapeutic applications will require good manufacturing practices (GMP) during both derivation and subsequent culture, including validation and documentation of manufacturing and testing protocols (see [27] for a recent review of GMP regulations as they pertain to hESC-based therapies). Since hESC lines can incorporate animal components, including the nonhuman sialic acid Neu5Gc, during culture in the presence of animal feeder cells or animal-derived serum components [28], deriving and culturing hESC lines in defined, humanized conditions would be ideal. However, lines that have been exposed to animal components may lose Neu5Gc following culture in humanized conditions [29,30], suggesting that contaminated lines may still be useful in clinical applications if they are proven to be free of infectious or pathogenic agents.

Allogeneic human fibroblasts, as well as hESC-derived cells, offer xeno-free feeder options for the derivation of hESC lines [3]. Recently, human feeder cells have been used in conjunction with medium containing human serum to derive xeno-free hESC lines [31,32]. Ludwig et al. described the first defined, completely animal-component free hESC derivation. Their medium contained a combination of small molecules and proteins including bFGF, TGFβ1, LiCl, GABA, and pipecolic acid and hESCs were cultured on a matrix consisting of human collagen IV, vitronectin, laminin, and fibronectin [33]. Although these lines acquired chromosomal abnormalities, including XXY and chromosome 12 trisomy, it remains unclear to what extent the defined culture conditions contributed to these abnormalities.

Sources of embryos for hESC derivation

The initial hESC lines were derived via harvest of the inner cell mass from blastocyst stage embryos and subsequent subculture on MEF feeder cells [1]. While most hESC derivation methods still utilize the entire inner cell mass of the blastocyst, lines can also be created from the morula or arrested embryos, albeit at a lower efficiency than from blastocysts [46,47]. Embryos with a low morphological grade do not efficiently yield hESC lines if they are arrested prior to the blastocyst stage (<1% efficiency), but at the blastocyst stage 8.5% of poor quality embryos formed hESC lines in one study [48]. Chung et al. derived multiple hESC lines using single cell biopsies from morula-stage embryo without affecting the ability of the biopsied embryos to develop to the blastocyst stage in vitro [49]. This technique, similar to that used in preimplantation genetic diagnosis (PGD), has not yet been used to generate an hESC line for an embryo that implanted and formed a complete organism. Given the labor-intensive nature of this process and the risk that blastomere biopsy may pose to the developing embryos, it is unlikely to become a widespread strategy for generating patient-specific hESC lines, but it may be beneficial for creating lines for embryos that undergo PGD.

Inducing fusion of murine embryonic stem cells with adult thymocytes reprogrammed the adult nucleus to an embryonic state [50]. Similar approaches successfully fused existing hESC lines with human foreskin fibroblasts and hESC-derived myeloid precursors to generate pluripotent cells that express hESC-specific genes and exhibit embryonic DNA methylation patterns [51,52]. The tetraploid nature of these cells makes them unsuitable for clinical applications unless the hESC chromosomes can be removed before or after fusion.

Revazova et al. demonstrated that blastocysts produced by chemically-induced parthenogenesis, asexual development of an unfertilized oocyte, can be used to derive hESC lines [53]. Parthenogenesis of human oocytes has also been used to generate human HLA homozygous hESC from HLA heterozygous donors [54]. While parthenogenesis results in a high incidence of genetic abnormalities (i.e. deleterious recessive mutations in the donor oocyte), this method does not require embryo destruction and may be suitable for generating HLA-matched lines which could be banked for patient-specific regenerative therapies.

Somatic cell nuclear transfer (SCNT) has been used to clone embryos and generate ESC lines from these embryos in a number of species, including primates [55]. A cloned human blastocyst-stage embryo was created by transfer of the nucleus from an undifferentiated hESC into a human oocyte [56]. More recently, SCNT of adult fibroblast nuclei into human oocytes was used to generate multiple cloned blastocysts with 23% efficiency. However, hESC lines have not yet been derived from these cloned blastocysts [57]. Hwang et al. reported, then retracted, generation of hESC lines from blastocysts formed by SCNT; interestingly, one of these lines was later demonstrated to be the result of parthenogenesis [58].

Induced Pluripotent Stem (iPS) Cells

A promising new source of patient-specific pluripotent cells is iPS cells (Table 1). These cells do not possess the same ethical concerns of hESC lines since they do not require the destruction of blastocysts. Yu et al. illustrated that primary and transformed human dermal fibroblasts can be reprogrammed to a pluripotent state by expression of OCT4, SOX2, NANOG, and LIN28 [59]. Simultaneously, Takahashi et al. demonstrated that four transcription factors, Oct4, Sox2, Klf4, and c-Myc, also induced pluripotency in human dermal fibroblasts [60]. MYC expression was not required to generate human iPS lines, but did improve efficiency [61]. Expression of hTERT and the SV40 large T antigen in addition to OCT4, SOX2, KLF4, and c-MYC also enhanced iPS cell generation from primary adult fibroblasts [62]. Human iPS cells express the hESC markers SSEA-3, SSEA-4, TRA-1-60, TRA-1-81, alkaline phosphatase, NANOG, OCT3/4, SOX2 [59,60]. Furthermore, iPS cells exhibit high telomerase activity, histone methylation patterns similar to those found in hESCs, and pluripotency as determined by embryoid body analysis, directed differentiation, and teratoma formation [59,60]. Global gene expression analysis of various iPS clones and hESC lines identified minor differences between iPS cells and hESCs and between different iPS clones, although these differences were no greater than those found between distinct hESC lines [59,60]. Further characterization of iPS lines, including assessing growth and differentiation potential, and the development of efficient induction factor delivery methods that reduce the heterogeneity of transgene expression will be required to establish the promise of these cells in cell-based therapeutics.