Introduction

Since the first publication in 1998 1, human embryonic stem cells (hESC) have attracted significant attention due to their dual ability to self-renew and to differentiate into all cell types of the body. This makes them excellent candidates for cell- and tissue-replacement therapies, as well as for the basic scientific research on human embryogenesis and diseases 2. However, most hESC lines available to date have been directly or indirectly exposed to animal material during their derivation and/or propagation in vitro 3, 4. Although there are efforts to use the currently available hESC lines for clinical trials, such xeno-contaminated cells are generally judged unsuitable for transplantation because of the risk of zoonosis transmitted by animal pathogens and the potential activation of animal retroviruses, not to mention the possibility of immune rejection 5. For therapeutic purposes, all steps for the production of hESC must avoid the use of animal components, i.e., culture systems using animal-free derivation methods, animal-free culture media, and animal-free substrates should be established to create and propagate clinical-grade hESC lines.

Derivation of hESC

hESC lines are conventionally derived from the inner cell mass (ICM) of pre-implantation stage blastocysts, of both good and poor quality 1, 6, 7, 8, which have been donated for research and would otherwise be discarded. Morula-stage embryos 9 or late-stage blastocysts (7-8 days) 10 may also be used to create hESC lines. Although all the hESC lines derived worldwide share the expression of characteristic pluripotency markers 11, 12, many differences are emerging between lines that may be more associated with the wide range of culture conditions in current use than with the inherent genetic variations of the embryos from which hESC were derived 13. The reported success rate for hESC derivation is highly variable, possibly depending on the embryo quality and the culture protocol used. Very recently, Klimanskaya et al. published a new method for the derivation of hESC lines from single blastomeres obtained by a procedure similar to the one used for preimplantation genetic diagnosis, which could, theoretically, avoid embryo destruction in the future 14. Alternatively, hESC lines were also successfully obtained from arrested embryos which stopped cleavage and failed to develop to morula and blastocysts in vitro 15. Such possibilities are expected to partially ease the ethical concerns surrounding the conventional hESC derivation techniques, which destroy viable human embryos. Further studies are needed to validate the efficacy of these new derivation protocols.

Isolation of ICM: immunosurgical versus mechanical methods

Isolation of the ICM from blastocysts is a very important step in the derivation of hESC lines. Although several reports demonstrated the derivation of hESC lines from intact blastocysts after removal of zona pellucida 16, 17, 18, the outgrowth of trophectoderm at an early stage might inhibit the expansion of ICM and further reduce the success rate of hESC derivation 19. Nearly all reported hESC lines to date have been efficiently obtained from ICM isolated by immunosurgery 20, a procedure that removes the outer trophectoderm epithelial cell layer from the blastocyst using anti-human whole-serum antibodies and guinea pig complement. However, the possible xeno-contamination in antiserum may limit its clinical use. Mechanical isolation of ICM serves therefore as a better alternative for the derivation of clinical-grade hESC lines, avoiding the exposure of blastocysts to animal antibodies. Human ESC lines have been successfully derived from manually isolated ICM 21. In our laboratory, laser-hatching technology is employed to mechanically isolate the ICM. Efficiently used in assisted reproductive medicine, laser beams can be precisely delivered to drill the trophectoderm and favor ICM expulsion. Compared with the manual microdissection, laser-hatching technology requires minimal micromanipulation skills and maximally avoids xenogenic contamination by reagents. In mice, ESC lines have been successfully established from the ICM isolated by laser hatching 22, and we are now adapting and optimizing such techniques to dissect ICM from human embryos in animal-free conditions. Other methods to obtain intact and healthy ICM, including enzymatic digestion as employed for porcine blastocysts 23, should also be considered for human embryos if proven effective using recombinant enzymes.

Feeder-dependent derivation of hESC

To obtain ESC, the isolated ICM need to be placed on specific substrates and cultured in appropriate media (Figure 1). In most cases, ICM are cultured on mitotically inactivated feeder cells. Although irradiated or treated with mitomycin, feeder cells are still able to stimulate ESC growth and inhibit their differentiation through the secretion of specific growth factors and cytokines 24. Like mouse ESC, hESC lines were firstly established on mouse embryonic fibroblasts (MEF) 1, in a medium containing fetal bovine serum (FBS). Such conditions are unsuitable for human cell transplantation since the exposure to animal components presents a serious risk of transmitting unidentified retroviruses and other pathogens to the patients. Although there is a report showing no evidence for hESC infection by animal feeder-derived viruses 25, concerns still remain over the clinical use of these cells. Furthermore, hESC cultured under xeno-contaminated conditions can present a non-human sialic acid, which is immunogenic to humans 5. For therapeutic purposes and to eliminate, or at least to reduce, the xenogenic contamination from animal feeder cells and sera, human feeder cells and serum replacement (SR) have been recently employed for both culture and derivation of hESC lines 17, 26, 27. One report also shows hESC lines established on feeder cells derived from hESC themselves 21. However, none of these culture systems can be considered entirely animal-free since FBS was normally used to culture feeder cells, and SR is a reconstituted formulation still containing large amounts of bovine serum albumin as well as other proteins. In order to completely remove animal substances from the culture system, several teams have employed human serum in culture media 16, 28. hESC lines derived under such conditions are expected to meet the criteria for clinical applications.

Figure 1
figure 1

Derivation and culture methods for human embryonic stem cells. Here it is illustrated all embryo stages showed in the literature to give rise to human embryonic stem cell lines. IVF: In-vitro fertilization; ICSI: Intra-cytoplasmic sperm injection; PN: Pronucleus.

Full size image

Feeder-free derivation of hESC

Like serum, feeder cells are a source of variability in experimental conditions and are also a concern for future transplantation of hESC derivatives to patients. A recent study successfully established a hESC line on the extracellular matrix (ECM) prepared from MEF 29. Despite the immunosurgical ICM isolation and the culture in SR-supplemented medium, the use of such ECM represents a significant advance as, if proven efficient, ECM of human origin would avoid the above-mentioned pathogenic risks, as well as the possible interference of living feeder cells in future transplantation therapies. In addition, feeder cell-derived ECM can be easily sterilized and stored for clinical applications. Another recent report 30 described the derivation of two hESC lines in a defined culture medium containing components solely derived from recombinant sources or purified from human material, on substrate formed by human ECM components. Although these culture conditions are laborious and costly, and may lead to abnormal karyotype during long-term culture, hESC lines derived and propagated in this xeno-free and feeder-independent system would be more directly applicable to clinical use. Such culture conditions also provide a basis for further simplifying culture requirements in studies investigating the molecular mechanisms of hESC self-renewal and differentiation. We are presently testing the capability of human feeder cell-derived ECM to support prolonged undifferentiated growth of hESC, aiming at the establishment of clinical-grade lines in suitably defined xeno-free culture conditions.

Clonal derivation of hESC

hESC derived from ICM of blastocysts should be considered as heterogenous 31. Isolation of a homogenous pool of differentiated hESC is necessary for basic studies like drug development and toxicity testing. Moreover, for therapeutic applications, the removal of undifferentiated, potentially tumorigenic cells will also be a requirement, or at least malfunctioning cells should be removed using cell ablation strategies, before and/or after engraftment.

New hESC lines have been clonally derived from the existing ones 32, 33, 34, 35. However, unlike mouse ESC, the clonal efficiency of hESC is extremely low, as hESC are sensitive to single-cell disaggregation and recover poorly when plated at clonal density. Indeed, cell-cell interactions seem critical for efficient hESC propagation, since the loss of gap junctions between hESC can increase cell apoptosis and inhibit hESC colony growth 36. For these reasons, hESC are routinely passaged in cell clumps, produced by mechanical cutting and/or mild enzymatic treatment 1, 8, which is an obstacle to large-scale propagation of hESC for clinical use. Nevertheless, clonal survival of hESC can be enhanced by culturing in physiologic oxygen (2%) 37, or in the presence of neurotrophins 38. Therefore, the efficiency of hESC derivation might be improved at low oxygen concentration and/or with neurotrophins in culture medium.

Passaging of hESC: mechanical versus enzymatic methods

When the feeder cells become old or the colonies are too dense or too large, hESC need to be split for continuing culture. Passaging of hESC falls into two categories – mechanical and enzymatic. Clearly, the passaging technique is critical for maintaining undifferentiated and karyotypically stable hESC lines. Mechanical passaging is performed by dissociating the colonies into small clumps through manually scraping and cutting and transferring them to new culture plates covered with feeder cells or appropriate substrates. This is labor-intensive and time-consuming, and it normally generates variable cluster size and results in inconsistent cell distribution. On the other hand, enzymatic passaging dissociates the hESC colonies using enzymes such as animal-derived collagenase, dispase, and trypsin 39, xeno-free recombinant enzyme 16 or enzyme-free cell dissociation buffer 40, allowing a relatively consistent and standardized passaging of hESC. In view of clinical applications, enzymatic passaging is advantageous since it enables large-scale expansion of hESC. However, emerging evidence suggests that excessive and/or frequent dissociation to single cells may lead to karyotype abnormalities 41, 42. In contrast, mechanical methods of passaging allow selective transfer of exclusively undifferentiated colonies and seem to better maintain genetic stability. Since the process of manual colony dissection limits the practical use of mechanical passaging in bulk hESC culture, many researchers combine both enzymatic and mechanical passaging methods for long-term hESC maintenance 8, 43. Indeed, hESC lines initially derived and passaged by mechanical methods have been subsequently adapted to enzymatic passaging for practical reasons 6. Automated mechanical passaging of hESC, which incorporates the advantages of mechanical dissection without sacrificing the practical benefits of enzymatic passaging, was also developed 44. To overcome the low resistance of hESC to single-cell dissociation during passaging, recent studies have achieved the culture of newly derived hESC lines which keep stable karyotype in long-term culture by enzymatic bulk passage 45, 46. These single-cell dissociation-resistant hESC lines would be valuable for clinical applications. Nevertheless, hESC in long-term culture need to be routinely checked for their chromosomal stability.

Propagation of hESC

Considerable progress has been made towards the definition of xeno-free culture conditions allowing the propagation of undifferentiated hESC, thus significantly improving the clinical potential of these cells.

Feeder-dependent propagation of hESC

To eliminate the xeno-contamination of mouse feeder cells, human feeder cells have been derived and shown to support the undifferentiated growth of hESC 28. The existing hESC lines can maintain their pluripotency on several types of human feeder cells 47, or on the feeder cells derived from hESC themselves 21, 48. However, most commonly used human feeders have been exposed to animal components during their isolation and culture 24, and the animal-derived SR is still needed to maintain hESC growth under these conditions, potentially making hESC xeno-contaminated and unsuitable for cell replacement therapies. In addition, hESC cultured in SR medium present immunogenic sialic acids on their surface; although the contamination with non-human sialic acids can be diluted (and probably lost) in hESC after long-term culture under xeno-free conditions, they may still represent a risk to transplantation 5. In view of clinical applications, it is mandatory to establish clinical-grade feeder cell lines for hESC derivation and propagation, while animal serum and SR should be eliminated from culture media for both feeder cells and hESC. Most recently, Ellerstrom et al. 16 established xeno-free human foreskin fibroblast feeders for hESC derivation, in a medium supplemented with human serum and devoid of any animal material. The hESC lines derived and propagated in such systems are therefore advantageous for possible therapeutic purposes.

Feeder-free propagation of hESC

The necessity of feeder cells limits the capacity to meet the large-scale culture demands for clinical applications, as well as for genetic manipulation of hESC in basic research. Furthermore, establishment of clinical-grade feeder cell lines for hESC culture is costly. Xu et al. 49 first demonstrated a successful feeder-free hESC culture system in which undifferentiated cells can be maintained in the long term on a Matrigel layer in MEF-conditioned medium. Alternatively, hESC-derived fibroblasts can also be used to produce conditioned medium capable of supporting undifferentiated growth of hESC on Matrigel 48, 50. However, Matrigel is a soluble basement membrane extract from mouse tumors containing several types of ECM and other unknown factors, and MEF-conditioned medium also contains undefined animal components, which limits its use for clinical-grade hESC culture. As a step forward, a single ECM component such as laminin or fibronectin, of both animal and human origin, has also been successfully used to support undifferentiated growth of hESC in either MEF-conditioned medium or medium supplemented with SR and various growth factors 49, 51, 52. In these conditions the possible xeno-contamination in substrates is reduced, but in culture media it remains and a risk to future clinical applications may still persist.

In order to optimize the culture conditions, great efforts have been made in identifying conditioned media components essential for hESC self-renewal, as a crucial prerequisite for the eventual applications of hESC in the treatment of human diseases. High throughput screening methods have been employed to investigate the protein composition of media conditioned by feeder cells of either animal or human origin 53, 54, providing preliminary insight into the possible feeder cell-secreted factors that support the growth of hESC. Basic fibroblast growth factor (bFGF) appears to play a key role in sustaining hESC self-renewal, and is included in nearly all the reported medium formulations for hESC derivation and propagation. Indeed, certain existing hESC lines have been successfully maintained over a long term in unconditioned medium supplemented with a high concentration of bFGF 55. Suppression of bone morphogenetic proteins (BMP) also seems to be important for hESC propagation. Noggin, a BMP antagonist, synergizes with bFGF in sustaining the proliferation of undifferentiated hESC in the absence of feeder cells and conditioned medium 56. Although the possible xeno-contamination of MEF-conditioned medium is avoided in these culture systems, exogenous factors in unconditioned medium (e.g. SR) and in culture substrates (e.g. Matrigel) still pose a risk to therapeutic applications. To circumvent this issue, Stojkovic et al. 57 reported the growth of undifferentiated hESC on human serum-coated plates in SR-containing medium conditioned by human embryonic fibroblasts derived from hESC, which reduces the exposure of hESC to animal ingredients. Recently, several chemical-defined culture media, in which the SR is replaced with a cocktail of recombinant growth factors and cytokines, have been developed to sustain undifferentiated propagation of existing hESC lines on Matrigel or on human-derived ECM in short-term studies 58, 59, 60. Such animal-free media should significantly facilitate the use of hESC in therapeutic applications. However, as discussed above, most available hESC cell lines are xeno-contaminated and their propagation in animal-free medium cannot completely eliminate the exogenous and immunogenic factors in culture 5. For clinical purposes, it would require starting with newly derived hESC that have never been exposed to animal products. As a step forward, Ludwig et al. 30 successfully derived and propagated two new hESC lines in defined medium on human-derived substrates, although immunosurgery is still employed in their work to isolate the ICM. Most recently, Fletcher et al. 61 reported the first hESC line derived without direct exposure to any undefined animal products. We expect such approaches to facilitate the clinical applications of hESC and to provide a platform for further optimizing culture requirements for undifferentiated growth of hESC.

It should be noted that, compared with the culture on feeder cells, the feeder-free system is less optimal for sustaining the undifferentiated growth of hESC, as differentiation on the edges of the colonies can frequently be seen in these conditions. Furthermore, feeder-free cultures of hESC might result in chromosomal changes, such as gain of chromosomes 17q and 12 62, which should be a concern for the future applications of these cells in transplantation therapies.

Other concerns

Despite the elimination of all animal components from the derivation, propagation and passaging process, there are still additional issues that should be addressed prior to the clinical applications of hESC. A general concern in all allogenic transplantations is the exposure of hESC to human feeder cells and human proteins in culture that may introduce unknown human pathogens into patients on cell transplantation therapy. Therefore, all the human substances used during the establishment and propagation of hESC must be extensively screened for the final therapeutic applications. Besides, as reviewed elsewhere 63, other roadblocks, such as tumorigenicity and genetic compatibility, also stand in the way of progress towards hESC-based cell therapies and these need to be removed before the clinical potential of hESC can be exploited.

Conclusion

To develop a clinical-grade hESC line, it is a prerequisite to show that all steps in the derivation, passaging and culturing of hESC are completely free of animal products. Ideally, such hESC lines should be derived from mechanically isolated ICM in a culture system which uses human-derived ECM and chemically defined medium, and which keeps normal karyotype for long term with xeno-free enzymatic passaging. Before the molecular mechanisms of hESC self-renewal are fully uncovered, it is currently acceptable to derive and culture hESC lines for possible therapeutic applications on human feeder cells in media supplemented with proteins that are solely of human origin.