Stem Cell Reviews Copyright © 2005 Humana Press Inc. All rights of any nature whatsoever are reserved. ISSN 1550-8943/05/1:87–98/$30.00
Original Article Human Blastocyst Culture and Derivation of Embryonic Stem Cell Lines Ariff Bongso* and Shawna Tan Department of Obstetrics and Gynaecology, National University of Singapore, Kent Ridge, Singapore 119074 Abstract Human embryonic stem cell (hESC) biology is expected to revolutionize the future of medicine by the provision of cell-based therapies for the treatment of a variety of deliberatig diseases. The tremendous versatility of hESCs has reinforced this hope. To understand the biology of these mysterious cells and attempt to differentiate them into desirable tissues, bona fide hESCs that maintain their stability with time are required for research and clinical application. This review discusses the various protocols to derive and propagate hESCs from high quality embryos. The nature and properties of hESCs are also described together with unanswered questions that need to be addressed if this science is to be taken to the bedside. Index Entries: Blastocyst culture; hESCs; derivation; properties; differentiation.
Introduction
*Correspondence and reprint requests to: Ariff Bongso Department of Obstetrics and Gynaecology, National University of Singapore, Kent Ridge, Singapore 119074. E-mail:
[email protected]
Stem cell biology has attracted much of interest recently. Daily, we read and hear news about how stem cells could revolutionize medicine and the promises it offers for the treatment of a variety of incurable diseases by transplantation therapy. Several types of stem cells have been isolated in the human and broadly they could be classified into two major types viz., embryonic and adult stem cells. Research on both stem cell types is equally important in order to achieve the ultimate goal of finding treatment for certain diseases. Although versatile in terms of pluripotentiality, research on human embryonic stem cells (hESCs) is politically charged, receives considerable media coverage, raises ethical and religious debates, and generates tremendous public interest because their source is from 5-d-old human embryos (blastocysts). In order to derive genuine hESC lines, good quality blastocysts are required. These are traditionally obtained from 1 to 3 d old frozen embryos left over after in vitro fertilization (IVF) treatment. In this review we discuss how human embryos could be scored, selected, and cultured in vitro to produce good
87
quality blastocysts. We then discuss the derivation and propagation of both research and clinical grade hESC lines from good and poor quality embryos. We conclude by describing the nature and properties of such derived hESCs. Most of the information in this review is our own work with interposing summaries of relevant areas of work carried out by other workers.
Human Embryo Culture Coculture and Sequential Media for Extended Human Embryo Culture Successful oocyte fertilization in vitro was first achieved in rabbits by Chang et al. in 1959. Later in 1969 Edwards et al. (1) accomplished this with human oocytes. This led to the birth of Louise Brown in 1978, and assisted reproduction techniques (ART) such as IVF and intracytoplasmic sperm injection (ICSI) have now become cornerstones in infertility treatment. The eventual success rate of IVF itself, often measured in terms of live birth rates, was <10% until the early 1990s (2,3). Other procedures, such as gamete intrafallopian transfer, yielded
88 ________________________________________________________________________________________________Bongso and Tan significantly higher rates (4,5) but these still fell well below the success rates achieved in domestic animals (>60%) (6). Much of this failure was attributed to poor sperm motility, oocyte chromosomal abnormalities, and suboptimal in vitro culture conditions (2). Previously, to ensure the best success rates, embryos were quickly returned to the uterus when they had developed to the 4- to 8-cell stage (days 2–3). If kept beyond this stage in vitro they would degenerate, get arrested, or die. However in vivo, 4- to 8-cell stage embryos reside in the fallopian tube instead of the uterus and hence in IVF programs embryos were being prematurely returned to an unphysiological environment viz., the uterus. It is at the blastocyst stage (day 5) that the human embryo reaches the uterine environment. Replacement of 4- to 6-cell stage embryos usually resulted in low-pregnancy rates, and optimal in vitro systems to culture human embryos to the blastocyst stage were not yet available. Interestingly, blastocyst-stage embryo replacement was responsible for a 60% success rate in domestic animals. Therefore an ideal culture system, which could replicate the human oviductal-uterine environment, was urgently necessary. This system would hopefully allow extended culture of human embryos and also improve embryo quality. When IVF culture conditions are suboptimal, human embryos undergo an in vitro embryonic block between days 2 and 3 because it has been reported that this is the time the embryo’s genome becomes activated (7). Previous IVF treatments necessitated the transfer of excessive quantities of embryos to overcome this problem, and subsequently had the negative result of multiple pregnancies (8,9). Culture procedures for embryos originally entailed using a single simple or complex culture medium for all embryonic stages, from the time of fertilization until blastocyst formation. This was disadvantageous as most of the resulting blastocysts were of inferior quality, fewer in quantity, and showed no improvement in implantation rates (10). Two experimental advances were thus made in embryo culture viz., coculture and sequential media. In our laboratory, human embryos at the 2 pronuclear stage (day 1) immediately after fertilization, were cocultured on a bed of human fallopian tubal epithelial cells plated in vitro, in the presence of a culture medium formulated along the composition of human fallopian tubal fluid. The advent of our coculture protocol allowed significant numbers of good quality blastocysts to develop in vitro (2,11–15). In our laboratory, human tubal ampullary cells also facilitated human embryo growth in vitro with 78% of embryos reaching the cleavage stage and 69% achieving cavitation, better than if they had been grown in a single standard culture medium (2). We also showed that among the different cell systems studied for their suitability in coculture, the sequential human oviductal–human endometrial in vitro system was most physiological and gave the best results (12–16). The pregnancy rates were doubled (42%) in a clinical trial using human tubal coculture in our center (11). We hypothesized that the expression of embryotrophic factors by positive conditioning and the removal of undesirable metabolites from the culture medium by negative conditioning were the modes of action for the positive coculture effects. The cells removed potentially harmful agents such as hypoxanthine and oxidation, which can impair early preembryo growth, and released the antioxidant taurine that was beneficial to embryonic growth. Several embryotrophic factors
such as interleukins and growth factors were released by the helper coculture cells and the culture medium was redesigned metabolically (15). Fukaya et al. (17) suggested that the mere physical contact of the embryo with the feeder cells may also bring about the advantages observed in coculture. Regardless of species, an embryo’s metabolic requirements change as its development progresses. A human embryo in the 2- to 8-cell stage has minimal metabolic activity and is unable to make full use of glucose; instead, it produces energy by oxidizing amino acids and pyruvate or lactate, and hence exhibit a greater requirement for these energy substrates compared to glucose. Embryonic growth progresses with the activation of its own genome, as well as the 2-cell types that morphologically identify it as an embryo: the inner cell mass (ICM) and the trophectoderm (TE). At this postcompaction stage (between days 3 and 5), the embryo’s metabolic activity increases, and the nutrients required to support oxidation, differentiation, and cell division shift to glucose, vitamins, nonessential, and essential amino acids (18–20). Genomic activation would have occurred and transcription begun, so that additional protein synthesis will consume the available amino acids. The sequential phases of embryo development tie in with the gradual changes that occur in the physiological environment of the female reproductive tract during natural pregnancy. Accordingly, we recommended the use of a more complex media formulation, with additional supplementation by vitamins and amino acids (21) to provide the most favorable external environment and allow the embryo to minimize the energy required to maintain its internal environment. This would enable it to overcome many of the metabolic blocks encountered in vitro, mimicking what happens in vivo. It is important to note that metabolic blocks can also occur during coculture owing to various reasons. During a natural pregnancy, Gardner et al. (18) reported that the highest levels of pyruvate and lactate are seen in the area of the oviduct nearest to the fimbrium, and these levels drop as the tract progresses, until they are at their lowest at the uterus. Conversely, glucose levels increase as the tract progresses from the oviduct toward the uterus. pH and hence oxygen and carbon dioxide levels also differ between the oviduct and uterine environments (22–24). The control of pH is critical to the developmental rate of the embryo. Embryos are known to regulate their intracellular pH in vitro between 7.2 and 7.4. High levels of amino acids, which regulate intracellular pH, are present in the female reproductive tract providing a pH-buffering action. The highest levels of hyaluronan are observed in the uterus and this aids selection of sperm at fertilization, cell division in the embryo, and its implantation. Hence, although routine culture of many other cell types utilizes only fixed culture conditions, it is necessary to meet the growing needs of the embryo by constantly providing it with a progressive culture system appropriate to its stage of development. Conventional human IVF culture media formulations now include salts, energy substrates, and serum or albumin protein sources. Serum albumin-supplemented Earle’s balanced salt medium is an example of a routinely used basic formulation containing the three basic components necessary for the growth support of early embryos (from fertilization to 4-cell stage) (21). Hardy et al. (25) reported that the bovine serum albumin-supplemented T6 culture medium, although
Stem Cell Reviews ♦ Volume 1, 2005
Human Blastocyst Culture and Derivation of ESC ___________________________________________________________________89 Table 1 Staging Human Embryos From Days 1 to 6 Day (am)
Embryonic stage
1
Two pronuclear (2PN)
2
4-cell stage
3
8-cell stage Compactinga Compacting Compacteda Early cavitatinga Late cavitating
4
5
Early blastocysta Expanding blastocysta 6
Fully expanded blastocyst Hatching blastocysta Hatched blastocysta
Description abutmenta
Pronuclear Nucleoli symmetry/alignmenta Cytoplasmic haloa 4 regular blastomeres, no fragmentsa 4 regular blastomeres, moderate fragmentsa 4 irregular blastomeres, many fragments 8 regular blastomeres Blastomeres compacting Blastomeres compacting Blastomeres compacted First signs of blastocoele Distinct blastocoele, ICM and TE not laid down Distinct ICM, TE, and blastocoele Embryo diameter same as day 4 or slightly larger Distinct ICM, TE, and blastocoele. Thin ZP Substantial increase in embryo diameter but not fully expanded Distinct ICM, TE, blastocoele. Thin ZP. Fully expanded diameter (approx 215 µm) ICM and TE hatching out of ZP ICM and TE completely hatched out from ZP
Abbreviations: ICM, inner cell mass; TE, trophectoderm; ZP, zona pellucida. a Ideal characteristics expected on that specific day.
highly capable of supporting the growth of embryos to the blastocyst stage, could not maintain it beyond this, resulting in the embryo expanding and then subsequently collapsing and deteriorating within 24 h. Other researchers have produced high-blastulation rates using Ham’s F12 or 10% human serumsupplemented Chatot Ziomek Barister (CZB) medium (26,27). The mechanisms behind the positive effects of coculture systems led to the development of cell-free sequential stagespecific culture media, since plating cells in vitro was labor intensive and there was the possible risk of cross-contaminating pathogens from the cells to the embryo. The original sequential media systems involved using one type of culture medium for culture from fertilization (day 1) to the 8-cell stage (day 3), and another type for growth to the blastocyst stage (day 5). Recent formulations mimic the female reproductive tract more closely, with human embryos grown in the first medium during days 1 and 2 and in a second medium during days 3–6. These sequential culture media differ in their composition. The first medium designed for days 1 and 2 human embryo is simpler with less glucose and more pyruvate than the second medium (days 3–5). The more complex second medium contains higher concentrations of glucose, amino acids, vitamins, ethylene diamine tetraacetic acid and taurine. Human embryo culture in sequential media is now widely used, has proven to be effective and has replaced coculture. Several companies now manufacture commercial sequential media formulations for human blastocyst culture.
Embryo Scoring and Selection The success of assisted reproductive procedures, in terms of high-pregnancy rates, depends on a noninvasive, nontraumatic evaluation of the preimplantation embryo by an experienced embryologist. To ensure that embryo transfers result
in pregnancies, the most viable embryos must be selected and this assessment has to occur within the first 5–6 d in vitro (28–34). Selection of good quality embryos also allows a reduction in the number of embryos to be transferred thus avoiding the occurrence of multiple pregnancies. Selection of the best embryos at different stages of embryonic development is carried out by morphological assessment. Using morphological parameters, implantation rates have increased to 10–30% for day 3 embryos, and 40–60% for day 5 blastocysts (35–38). This success has occurred in concert with technological advancements in oocyte insemination, micromanipulation and embryo transfer techniques, preimplantation genetic diagnosis (PGD), and improvements in culture media formulations. Morphological embryo scoring protocols were also useful for producing high-quality blastocysts for ESC production in our laboratory (39). We carry out our morphological evaluations on a warmed stage (37°C) under Hoffman’s inverted optics. The following are the morphological scoring criteria we use for each day of embryonic development (Table 1; Fig. 1).
Day 0 (Polar Body Morphology) Ebner et al. (40) showed that the morphology of the first polar body was strongly correlated with the quality of ensuing embryos on days 2 and 3 and final clinical pregnancy rates. They classified polar bodies into four types: smooth, ovoid, rough, fragmented, and large. The best oocytes had smooth ovoid polar bodies whereas the worst had large, rounded polar bodies.
Day 1 (Pronuclear Morphology and First Cleavage) In 1990, Van Blerkom (41) demonstrated that pregnancy rates were affected by pronuclei symmetry, alignment, and quantity of its nucleoli, as well as the appearance of the cytoplasm.
Stem Cell Reviews ♦ Volume 1, 2005
90 ________________________________________________________________________________________________Bongso and Tan
Fig. 1. Human embryonic stages in vitro. (A) Day 3 compacting embryo (B) Day 4 compacted embryo (C) Day 4 early cavitating embryo (D) Day 4 late cavitating embryo (E) Day 5 early blastocyst showing inner cell mass (ICM), trophectoderm (TE), and zona pellucida (ZP) (F) Day 5 expanding blastocyst showing ICM,TE, and thinner ZP.
A normal pronuclear embryo’s cytoplasm appeared heterogeneous with a clear crescent-shaped area (the cortical halo effect) in the embryo periphery, and a peripronuclear halo and dense cytoplasmic area surrounding the pronuclei (31,32,42). These cytoplasmic markers indicate that zygote polarization has occurred, whereas rearrangement of cellular components such as organelles will control the embryo’s subsequent development. Correct polarization is important as the second polar body must be positioned properly along the long axis, to ensure that the appropriate biological function is assigned to each zygotic pole. An incorrect distribution of the pronuclei and polar bodies is associated with cleavage errors that cause disproportionate blastomeres and decreased cleavage rate (29), and such embryos are poorer in quality and develop slowly. A standardized system of pronuclear scoring, in which good quality embryos could be selected to improve implantation rates, was devised by several researchers, such as Scott
et al. (42) and Tesarik et al. (32). Now known as zygote grading or pronuclear scoring, the assessment of the embryo starts within 16 to 18 h of IVF or ICSI procedures. Smith (43) also defined three parameters, which if present in combination, yield the best blastocysts. These were pronuclear abutment, symmetrical nucleoli alignment in rows, and a cytoplasmic halo around the pronuclei. Tesarik et al. (32) refined this and reported six morphologically different pronuclear embryos whereas nucleoli symmetry was given importance irrespective of nucleolus size. The best pronuclear embryos that generated good blastocysts were those in which nucleoli in both pronuclei were evenly scattered or aligned in rows. Another important scoring criterion on day 1 was the time the first cleavage occurred (30,34,44,45). Good quality embryos that will generate blastocysts cleave to 2-cells at 24–35 h postinsemination and will exhibit little fragmentation and blastomeres that are equivalent in size and symmetry.
Stem Cell Reviews ♦ Volume 1, 2005
Human Blastocyst Culture and Derivation of ESC ___________________________________________________________________91
Day 2 (4-Cell Stage) Embryos on day 2 at the 4-cell stage are classified morphologically into three types (good, fair, and poor) in our laboratory. Good day 2 embryos have regular, equal-sized blastomeres, with no cytoplasmic fragments, whereas fair embryos accumulate modest amount of fragments, and poor embryos have many fragments and irregularly sized blastomeres.
Day 3 (8-Cell to Compacting Stages) A day 3 embryo has a higher chance of implantation success than a day 2 embryo, once the in vitro embryonic block is bypassed and the embryonic genome is activated. We observed in our laboratory two distinct morphological stages on day 3 for good quality human embryos that finally produce blastocysts. These are stages with eight equal-sized blastomeres and no fragments, and the compacting stage, in which the blastomeres begin to compact closely with one another. The compacting stage shows faster cleavage and receives a higher score than the 8-cell stage.
Day 4 (Compacted and Cavitating Stage) On the morning of day 4, the best embryonic stages that usually result in blastocyst production are the fully compacted and cavitating stages in our laboratory. In the fully compacted stages, all blastomeres have compacted with one another, representing the classical “morula” stage described in some texts. We have observed that the cavitation stage is a faster cleaving and the better day 4 stage. In this stage compaction has completed and the first signs of the blastocoele have begun to appear.
Day 5 (Blastocyst Stage) Assessment and grading of the blastocyst was developed when it became possible to integrate embryo transfer on day 5 with extended culture techniques (35,37,42,46). Gardner (47) and Fong and Bongso (48) from our laboratory proposed fairly similar blastocyst scoring systems. Three types of blastocysts could be observed on days 5 and 6: (1) early blastocysts, (2) expanding blastocysts, and (3) fully expanded blastocysts. The ICM, TE, and blastocoelic cavity are laid down in all three types and hence the embryo is defined as a blastocyst. The major difference between the three types is the increase in diameter of the embryo, resulting from the accumulation of blastocoelic fluid and the thinning of the zona pellucida (ZP). Expansion of the blastocyst was correlated with cleavage speed, and the faster growing expanding blastocysts were of better quality as confirmed by increased total cell numbers (TCN) (48) (Table 2). We showed that TCNs increased from expanded to hatching stages in both coculture and sequential culture media (48). Additionally, the size of the ICM is the singlemost important criterion of blastocyst quality. When the ICM extended to at least half the diameter of the embryo the blastocyst was graded the highest quality (Table 2). This is also in line with the derivation of hESC lines because the success rates of immunosurgery and subsequent hESC line derivation is quite high with larger ICMs in blastocysts.
Embryo Cryopreservation In ART programs, ovarian stimulation sometimes leads to an excessive number of oocytes. Since oocytes cannot be
Table 2 Markers for Blastocyst Viability Morphological Characteristics ICM size Thin ZP Single large blastocoele “Sickle-shape” in cells in TE Cleavage speed Cavitated, expanding—day 5 (am) Fully expanded—day 5 (pm) Hatching—day 6 (am) Abbreviations: ICM, inner cell mass; TE, trophectoderm; ZP, zona pellucida.
frozen, they all have to be inseminated and as a result more than the required number of embryos for transfer are produced. Such surplus embryos are cryopreserved at the 2-pronuclear, cleavage, or blastocyst stages for future replacement into the patient, donation or for research by informed consent. Pronuclear and early cleavage stage embryos are frozen, with 1, 2-propanediol (PrOH) (49) or dimethylsulfoxide (50,51) as cryoprotectants using the slow programmable machine freezing method. The choice of cryoprotectants and freezing method depends on the developmental stage of the embryo. Embryos are loaded into freezing straws and the straws are sealed at both ends and then cooled in the freezing machine from room temperature to –7°C at a rate of –2°C/min. The straws are held at –7°C for 5 min for manual seeding to induce ice crystal formation, then cooled to –30°C at 0.3°C/min and to –150°C at 15°C/min before plunging into liquid nitrogen (LN2) (–196°C). Other methods, which are still experimental but used in some ART centers, include ultrarapid freezing and vitrification. Vitrification is a form of cryopreservation, in which instant cooling to –196°C is achieved by inducing glass instead of ice crystal formation. Blastocysts are frozen using two cryoprotectants (5% and 9% glycerol) and the slow programmable machine freezing method. Thawing of 2-pronuclear, cleavage, and blastocyst stage frozen embryos is done by slow stepwise removal of the cryoprotectants into a final growth culture medium. Traditionally, embryos donated for deriving hESC lines are frozen at early cleavage stages on days 2 or 3. Such embryos have to be thawed and then grown in fresh sequential culture media to the blastocyst stage for hESC derivation. Alternatively, commercial freeze/thaw kits are useful in the freezing and thawing of human embryos for hESC derivation.
Derivation and Propagation of Research Grade hESC Lines From Good and Poor Quality Human Embryos Since some hESC lines exhibit phenotypic discordance and thus their usefulness to research can vary significantly, it is important to continue to derive new and additional cell lines that are stable and fully characterized. The storage of panels of HLA typed hESC lines containing information with respect to growth characteristics, phenotype, genotype, culture conditions, gender, and stability during prolonged passage, will be very useful for research and application.
Stem Cell Reviews ♦ Volume 1, 2005
92 ________________________________________________________________________________________________Bongso and Tan
Derivation of hESC Lines From Surplus IVF Embryos hESC lines can be derived from embryos donated for research, after informed patient consent and approval from institutional review boards. Such embryos are usually surplus embryos generated during IVF treatment cycles. Pregnancy rates are generally poor when only a single fertilized oocyte is produced for subsequent embryo transfer from a natural cycle. To obtain the best success, at least 2–3 embryos are usually transferred. Therefore a subfertile female patient has to receive hormonal ovarian stimulation to produce at least 8–10 oocytes so as to generate a minimum of 2–3 embryos given the fact that fertilization and cleavage rates are 60% and 80%, respectively. Thus in some IVF patients surplus embryos are generated. These are either donated to (1) another couple seeking infertility treatment, (2) for research, or (3) disposed. Embryos donated by patients for informed research are the major source of material for deriving hESC lines.
Propagation by the Mechanical Dissection Method hESCs were first isolated in 1994 from 21 IVF blastocysts donated by nine IVF patients in our laboratory (52). We used the whole embryo culture method in which blastocysts were first incubated with pronase to digest the ZP, and the intact ICM and TE were plated on γ-irradiated, mitotically inactive human adult oviductal epithelial fibroblast feeders, in Chang’s medium containing 1000 IU/mL hLIF. After 7–11 d in culture, these zona-free blastocysts developed into ICM clumps and were removed from adjacent trophectodermal and feeder cells, by mechanical dissection with hypodermic needles. We enzymatically disaggregated the clumps with trypsin before seeding onto freshly irradiated human adult fallopian tube cells. hESC colonies formed at the first two passages and exhibited characteristic hESC features such as alkaline phosphatase expression and hESC morphology (distinct and prominent nucleoli; increased nuclear:cytoplasmic ratios). We also showed that they were karyotypically normal. They subsequently differentiated after the third passage (15,52). Later, Thomson et al. (53) derived the first hESC line by separating the ICM using immunosurgery and plating the ICM on irradiated mouse embryonic fibroblast (MEF) feeders instead of human feeders. Dulbecco’s modified Eagle’s medium with high serum concentrations was used as the culture medium and these ICMs generated hESC colonies which were transferred to freshly inactivated MEFs by mechanical dissection, rather than trypsinization. By repeated mechanical dissection in subsequent passages, a hESC line was established. It was perhaps the disruption of the junctional complexes between the hESCs with trypsin that may have hindered the production of hESC lines in our laboratories (15,52). A potential shortcoming of immunosurgery is the exposure of the blastocysts to animal antibodies, which can negatively impact future developments in clinical settings. Mechanical separation of the TE carries its own risk as some TE cells can be left behind, which could outgrow the ICM cells and inhibit them (54). Reubinoff et al. (55) later devised a protocol combining immunosurgery, mitomycin-C inactivated MEFs, and mechanical dissection to derive and grow hESC lines that could be maintained for prolonged passages and which could be spontaneously induced to differentiate into neurons. The
“whole embryo culture” method initially developed in our laboratory (15) together with mechanical dissection was recently used to develop hESC lines (56–58). More recently, human feeders from fetal, neonatal, and adult sources have been used to derive and grow hESC lines by us and other workers (58–61). hESC lines for research have also been derived using abnormal, mononuclear zygotes that are not suitable for embryo transfer in IVF programs (57) and also from normal but poor quality embryos (62) which would otherwise have been destroyed. The latter were derived by the same approach as Thomson et al. (53), using immunosurgery to isolate the ICM and subsequent culture on MEFs. This group also passaged hESC colonies via mechanical dissection. Although the mechanical disassociation method is suitable for efficiently propagating undifferentiated hESC growth in vitro over prolonged periods, it is limited in terms of scaling up cell numbers and the quantity of cells that can be grown, as it requires significant labour, time, and a highly trained eye to recognize undifferentiated colonies. An alternative would be to increase the scale of cell growth through enzymatic bulk culture methods.
Propagation by the Enzymatic “Bulk Culture” Method Culturing hESCs in bulk allows colonies to be scaled up in large numbers in a comparatively short space of time. To initiate such a culture, undifferentiated regions of selected hESC colonies are first mechanically dissected into fragments and replated on mitomycin-C inactivated MEFs. After several early passages, the colonies are dissociated into small, individual clusters by incubation with collagenase and/or trypsin, before being pooled together and seeded as a suspension onto freshly mitotically treated MEFs. This allows numerous colonies to form much more quickly than by mechanical dissection. Variations of this method exist in different laboratories, such as the simultaneous enzymatic dissociation of all the cells (hESC colonies and feeders), and then plating the mixed pool of cells onto new feeders. During culture, the hESCs will survive and proliferate to generate new colonies, whereas the remaining feeder cells will degenerate and not grow further. In our laboratory human feeders have replaced mouse feeders for bulk production of hESCs (Richards and Bongso, unpublished data). Cowan et al. (63) used 97 ICMs to derive and characterize 17 hESC lines on MEF feeders in serum replacement-supplemented culture media containing recombinant hLIF and plasmanate. After outgrowths were produced from the ICM, mechanical cutting was employed to propagate new colonies for the first five passages and subsequent passages were carried out with trypsin. A normal karyotype was present in all cells in culture, but changes were later detected in chromosomes 2 and 12 after prolonged hESC serial culture. These abnormalities may confer certain growth advantages on the cells, and allow a reduction in the culture time required for doubling.
Derivation of hESC Lines From Embryos Created by Nuclear Transfer To customize hESCs to patients to prevent rejection during transplantation therapy, an alternative approach to the derivation of hESCs is via nuclear transfer (NT) on embryos created for this purpose. Although Li et al. (64) demonstrated that
Stem Cell Reviews ♦ Volume 1, 2005
Human Blastocyst Culture and Derivation of ESC ___________________________________________________________________93 undifferentiated and differentiated hESCs injected into immunocompetent mice do not elicit an immune response, and are hence immunoprivileged, there are still concerns that the use of IVF embryos to derive hESC lines will cause transplantation incompatibility and rejection. To circumvent this, Hwang et al. (65) employed somatic cell NT (SCNT) to generate cloned embryos and derive a hESC line. They used cumulus cells and 176 enucleated oocytes from a single patient for this procedure. Blastocysts were then created by electrofusing the cumulus nuclei with the enucleated mature oocytes. Subsequent isolation of the ICM was by immunosurgery, plating on MEFs, and establishment of a hESC line using the mechanical dissection method. However Trounson (66) stated that since SCNT required large quantities of oocytes in order to produce just one hESC line, it would be impractical at this moment in time to apply this process to every patient who requires customized hESC-derived therapy. Furthermore, Daniels et al. (67) suggested that other scientific issues need to be addressed, including the consequence of the lack of spermmediated gene imprinting, the existence of mitochondrial DNA remnants in the oocytes after enucleation, the lack of mature oocyte availability, and also, the possible incidence of developmental anomalies from epigenetic remains of the donor’s somatic cell.
Success Rates of hESC Derivation Factors for successful hESC derivation include the availability of adequate quantities of good quality embryos and having an efficient culture procedure to support blastocyst development. Although the derivation of new hESC lines is no longer novel, it is still difficult to determine how successful different researchers have been with the protocols because of the differences in the way they have been documented. For example, in some groups’ report the quantities of donated cleavage-stage embryos rather than blastocysts, and the quality of the blastocysts used are not described. A rough idea of success rates in hESC derivation can be obtained from published work in which the numbers of embryos and hESC lines were given. Thomson et al. (53) derived five hESC lines from 14 blastocysts, Reubinoff et al. (55) derived two lines from four blastocysts, Pickering et al. (54) derived three lines from 58 embryos and Cowan et al. (63) reported the successful derivation of 17 lines from 97 ICMs, of which 286 were isolated from early embryos and 58 from blastocysts. Four hESC lines were developed from 10 blastocysts in our laboratory (Fong and Bongso, unpublished data). Thus, success rates can extend from 20% to 40% depending on the skills and quality of blastocysts used. As the learning curve and more experience is developed the success rates will go up for the derivation of hESC lines from surplus IVF embryos.
PGD and Derivation of Genetically Abnormal hESC Lines Embryo selection is an important contributory factor toward the success of ART. At present, this is best achieved using noninvasive morphological characteristics and a scoring system for each day of embryonic development (days 0–5) in vitro. A high cumulative or graduated embryo score helps in identifying the best embryos for transfer into IVF patients. Traditionally,
the remaining surplus average-to-poor quality embryos are cryopreserved. Some ART centers cryopreserve the best quality embryos and transfer the poor quality embryos. Munne et al. (68) showed that poor quality embryos are often aneuploid although some good quality embryos may also contain chromosomal anomalies. As such, PGD has become a useful tool to screen human embryos before transfer. The method has also been used routinely in some ART programs to screen thalassaemias, age-related aneuploidies, sex-linked and autosomal genetic diseases. One to two blastomeres are aspirated at the 8-cell (day 3) stage using micromanipulation and specifically designed aspiration micropipettes. The ZP is thinned enzymatically or with acid-Tyrode’s solution prior to insertion of the micropipet into the embryo for blastomere aspiration. The 1–2 blastomeres are then subjected to fluorescent in situ hybridization, using multicolour multiprobes for at least seven known genetic diseases. The remaining 6/8 or 7/8 embryo is viable and can be either frozen or transferred fresh into the patient once a genetic diagnosis is made. Once an embryo is diagnosed as genetically abnormal, consent can be requested from the donor of that abnormal embryo for hESC derivation. Such material becomes a very useful resource for generating banks of genetically abnormal hESC lines for studies in the future. Patients evaluated by PGD for single gene or X-linked disorders or translocations were asked by Pickering et al. (54) to consider donation of their embryos for hESC line derivation if they were at higher risks for genetic disease or had a genetic disorder after PGD. Eight patients going through PGD gave consent, with five for sex-linked disorders (36 embryos), one for spinal muscular atrophy type I (two embryos), and two for chromosomal translocations (six embryos). A total of 44 embryos were cultured from days 4 to 5 or 6 and 24 embryos (55%) developed to blastocysts. Two embryos generated two hESC lines with bonafide looking hESCs containing genetic defects. One of these lines was derived from a male embryo donated after a PGD cycle for an X-linked disorder (Becker’s muscular dystrophy), and the other from an embryo identified at higher risk for SMA type I. The two hESC lines were fully characterized. Such hESC lines become valuable tools to investigate genetic disease progression and efficacy and toxicity of drug therapies at cellular and molecular levels, before and after differentiation into specific cell types (54). Such genetically abnormal hESC lines also become useful in comparative transcriptome profiling with genetically normal hESC lines in search of undifferentiation and differentiation factors.
Derivation and Propagation of Clinical Grade hESC Lines In the US, in order to receive federal funding for research, ESC lines must be registered with the National Institutes of Health (NIH). To date however, all of the 78 hESC lines registered with NIH have had some physical contact with a xeno (animal) component. They have been derived and cultured on animal feeder cell supports such as MEFs or animal-derived feeder cell-free matrices and/or culture media supplemented with animal proteins (e.g., fetal calf serum [FCS], bovine insulin, and porcine transferin). Additionally, they have also been exposed to animal based products during the derivation procedure such
Stem Cell Reviews ♦ Volume 1, 2005
94 ________________________________________________________________________________________________Bongso and Tan as guinea pig complement for separating the ICM and bovine/ovine hyaluronidase for ICSI. Such protocols exponentially increase the exposure of the cells to possible animal pathogens such as the Hantavirus and lymphocytic choriomeningitis virus from MEFs and their conditioned medium, and can potentially contaminate a differentiated hESC line carrying remnant animal feeder cells. To be adapted for clinical use, we suggested that hESCs need to be derived and grown using xeno-free protocols and current good manufacturing practices (cGMP) as well as good tissue culture practices (69). Alternatives to the use of animal feeder cells are the use of nonanimal support matrices or the substitution with infectionfree human feeders for the derivation and culture of hESC lines (59,70–72). Unfortunately protocols using cell-free support matrices still have some degree of animal exposure because the majority of available matrices are animal-derived, and/or the culture media used with such matrices are conditioned with MEFs. We also believe that at this point in time it is virtually impossible to plate ICMs on cell-free matrices and derive hESC lines although propagation of already derived hESCs on matrices is possible. Thus screened human feeders in the presence of human-based culture ingredients in the culture medium appear to be the best choice at present for both derivation and propagation. Work in our laboratory showed that feeders derived from fetal sources are typically better at supporting hESCs compared to human adult sources (59,60). We also showed that human adult skin taken as a small biopsy from the abdomen, expanded in culture, mitomycin-C treated, and then used as a feeder also supported quite well the growth of hESCs without significant differentiation. Thus, one approach would be to obtain from the same patient donating embryos for hESC derivation a skin biopsy for producing feeders to support her own hESC growth. This would create an in vitro autologous rejection-free system, and such feeders need not be screened as the patient donating them would have already been screened for HIV, hepatitis B, and other diseases before being enrolled for IVF. We have also successfully derived and propagated hESCs on human fetal muscle feeders in culture medium supplemented with either human components (serum, insulin, and transferrin) or the proprietary knockout serum (59). Studies in our laboratory on the growth of hESCs on a panel of human fetal and adult feeders (60) showed that human feeders can be classified as supportive and nonsupportive and the top three supportive feeders in terms of best hESC growth were human fetal muscle, human fetal skin, and human adult skin in that order. Human fetal skin feeders were obtained from commercial sources (D551, ATCC, Maryland) whereas the rest were derived in-house in our laboratory as previously described (73). Later, Hovatta et al. (61) and Amit et al. (71) also successfully used human foreskin fibroblasts to sustain the prolonged culture of hESCs. However, many commercially available cells that can be used as feeders are not ideal, as they have been in contact with animal components like FCS. The most ideal system is one that is feeder-free and xeno-free right from the start, from derivation through to hESC serial culture. This would expedite and ease the scaling up of cells for research and clinical application, and avoid the risk of contamination with pathogens. A thorough transcriptome profile that reveals the genetic composition and protein expression of hESCs may
allow in the future the establishment of hESC lines without the need for feeder support. Instead a xeno-free culture medium containing the proteins/factors expressed by the stem cell signature genes could be used.
Xeno-Free Cryopreservation of hESCS In line with the aims to create therapeutically acceptable, cGMP-compliant and xeno-free hESC lines, a safe and effective cryopreservation system must also be available. Such new protocols should eliminate exposure to LN2 since it has been shown that some adventitious agents survive in LN2 and could be transmitted to the hESCs if the cells are buried in the liquid phase of LN2 (74–77). The majority of stem cell laboratories currently cryopreserves hESCs in closed cryovials by conventional slow-freezing and stores them in the liquid phase of LN2. The cryoprotectant used is a mixture of 90% FCS and 10% dimethylsulfoxide. The thaw-survival rates are very poor and it has been suggested that the induction of ice crystal formation during freezing causes intercellular disruptions. An alternative ultrarapid snap-freezing (vitrification) method that works on the principle of glass induction instead of ice crystal formation is being used by some workers. This method however uses straws open at one end, open pulled straws (OPS), for vitrification with ethylene glycol as cryoprotectant (78). Both the slow freezing and OPS vitrification methods have drawbacks. Although much simpler than slow freezing, the opening at one end of the straw in the vitrification method causes LN2 to come into direct contact with the hESCs when immersed in the liquid phase of LN2. The Hepatitis B and HIV viruses, among others, can survive in LN2 (76,77) and thus provide possible contamination of frozen stock. With slow-freezing, hESC survival and undifferentiated growth rates after serial passaging are very low. Additionally, in the slow-freezing method, when cryovials are stored in LN2 a vacuum forms inside the tube owing to condensation of the atmosphere and this can cause LN2 to seep in, compromise sterility, and burst during thawing, causing potential harm to the operator. We recently reported the development of a safe, xeno-free and effective cryopreservation method utilizing closed-straw (CS) vitrification and storage in the vapor phase of LN2 (79). To prevent contamination, the hESCs were drawn into conventional IVF “embryo straws” and heat-sealed at both ends before storage in the vapor phase of LN2 instead of the liquid phase. Additionally, we replaced FCS with human serum albumin (HSA) in the cryoprotectant thus making the system xenofree. This cost-effective, xeno-free, and highly reproducible method relies also on glass formation brought about by the rapid cooling of the cells in the straw with cooled tweezers. Other problems such as dehydration, ice crystal formation, and the need to carefully regulate freezing rates are avoided, as well. We showed that hESCs thawed and cultured after CS vitrification yielded high-undifferentiated rates (88%) and maintained pluripotency marker expression, teratoma formation in immunodeficient mice and normal karyotype. Work is in progress in our laboratory to adapt this new freezing method to hESCs propagated by the expandable bulk culture method.
Nature and Properties of hESCs A true hESC can be defined by a variety of unique intrinsic morphological and functional traits. For example, unlike
Stem Cell Reviews ♦ Volume 1, 2005
Human Blastocyst Culture and Derivation of ESC ___________________________________________________________________95
Fig. 2. Mophology of hESCs. (A) hESC colony grown on human fetal muscle fibroblasts by the mechanical dissection method. Note flat, thin colony boundaries, phase contrast 100X. (B) Stereomicrograph of hESC colonies grown on human fetal skin fibroblasts by the mechanical dissection method, 60X. (C) hESC colony grown on human fetal skin fibroblasts by the enzymatic bulk culture method. Note that colony boundaries are not defined, phase contrast 100X. (D) hESC colony grown on MEFs by the mechanical dissection method. Note thick, circular colony boundaries, phase contrast 100X (E) stereomicrograph of hESC colony grown on MEFs by the mechanical dissection method 60X. (F) High magnification of hESCs showing prominent nucleoli and high-nuclear-cytoplasmic ratios, phase contrast 200X.
somatic cells, hESCs have the ability to develop into all adult cell types via the three primordial germ layers even after extended culture in vitro, that is, they are pluripotent. On differentiation, these characteristics gradually change or are replaced by different ones. Thus bonafide hESCs have a very unique nature and set of properties.
Morphological Behavior Adult somatic cells do not persist indefinitely in culture and will senesce after a period of growth in vitro. On the other hand, hESCs in culture can survive indefinitely, thus classifying them as immortal by some workers. These cells are unique in morphology and at high magnification appear to have small quantities of cytoplasm surrounding a large nucleus with highly conspicuous nucleoli thus giving them high nucleus to cytoplasm ratios (Fig. 2). The method of hESC culture also influences hESC growth and colony morphology. In the mechanical dissection method using MEFs, more compact circular colonies with thicker borders are observed (Fig. 2). In contrast, hESC colonies propagated on human feeders display a more angular and rhomboid morphology, with a thinner periphery (Fig. 2). Colonies grown on MEFs also have a greater tendency to differentiate earlier than those cultured on human feeders, and usually have to be passaged by 7 d compared with 8 d for those grown on human
feeders. If left in culture beyond this, differentiation tends to occur in the center and periphery of the colony and further passaging can be carried out by careful dissection of the undifferentiated parts avoiding the center and periphery. In the bulk-culture method, hESC colonies are more numerous but small with no defined peripheries. The time taken for hESC populations to double depends on the stage of growth. Generally, we have noticed that population doubling occurs every 24–96 h, but cells in the earlier passages take a significantly longer time to double. After serial passages, the cultures adapt to in vitro conditions and become more stable in their growth. Amit et al. (80) showed that the clonal expansion of single hESCs is possible but inefficient owing to difficulties related to the propagation of a single cell. However, by dissociating hESCs into clumps and propagating them in suspension, it is possible to induce the formation of embryoid bodies (EBs), which are capable of differentiation into a range of cell lineages, for example, cardiomyocytes or neurons. This is a feature unique to hESCs and not observed in adult somatic cells.
Chromosomal Stability Although hESCs have a stable developmental capacity, there is concern that prolonged cultivation in vitro can result in karyotypic changes. Such changes may confer selective growth advantages in vitro but preclude their use in the development
Stem Cell Reviews ♦ Volume 1, 2005
96 ________________________________________________________________________________________________Bongso and Tan of clinical therapies. The majority of NIH-approved cell lines have normal karyotypes. However chromosome changes could occur during culture and routine karyotyping of cells at the point of derivation or early culture, and at regular points thereafter, should be a necessary part of stem cell culture procedures. Draper et al. (81) reported chromosomal anomalies in two hESC lines, H7.S6 and H14.S9 (University of Wisconsin). These were an amplification of chromosomes 17q and 12p in hESCs that were cultured for extended periods of several months (22–60 passages). The cells produced trisomies for chromosome 17q, and the entire long arm of chromosome 17 was translocated to chromosome 6. Similar gains of chromosome 12 as a result of amplifications of its short arm were also noted. The authors suggest that culture methods which favor the growth of undifferentiated hESCs may weed out mutations that cause spontaneous differentiation and cell death. One of the traits of malignant cells is their immortality, and it is interesting to note that these particular chromosomal aneuploidies are similar to those detected in human embryonal carcinoma cells, which have one extra 17q and one or more 12p isochromosomes. It is highly likely therefore, that these chromosomes have important roles in self-renewal and these particular aberrations may facilitate stem cell proliferation. The cells may also have adapted, under the conditions or manipulations during culture in vitro, and selected for genes that will help them thrive (82). Such environmental factors include the use of feeder-free cultures, or the passaging of cells at high densities. In contrast, in a separate study, Buzzard et al. (83) did not observe chromosomal abnormalities in five other lines (ESI, HES1-4, and HES6), which were cultured for the same number of passages and propagated by the mechanical-dissection method. They suggested that the mechanical-dissection method of passaging, rather than enzymatic dissociation, may have contributed to the cytogenetic stability. The enzymatic bulkculture method was associated with similar karyotypic anomalies observed in the hESC lines of another group (63). Similarly, three different hESC lines, HS181, HS235, and HS237, initially had normal karyotypes at passages 39, 39, and 35, respectively, but when HS237 was retested at passage 61, an isodicentric Xchromosome anomaly was detected (84). Chromosomal changes may not adversely affect the development of clinical therapies if large numbers of cytogenetically normal, early passaged cells can be generated.
Transcriptome Analyses Athorough understanding of gene expression in hESCs and their differentiated counterparts will shed light on the undifferentiation and differentiation mechanisms and help to direct hESCs to stable lineages. It is very important that the regulatory mechanisms involved in pluripotency and differentiation processes is properly understood. Various strategies have been employed to characterize the hESC transcriptome and elucidate the genetic control pathways that maintain pluripotency and differentiation, including microarrays, serial analysis of gene expression (SAGE) and expressed sequence tags (ESTs). SAGE is a useful procedure for the detection of short and long tag sequences (short SAGE and long SAGE) as it helps to identify novel gene sequences. In an effort to define the molecular mechanisms behind “stemness,” hESC self-renewal and differentiation, we studied
two undifferentiated hESC lines using SAGE (85). Of 145,015 SAGE tags that we produced, almost 21,000 were considered unique and many of which could be matched to novel genes, ESTs and hypothetical proteins. Some of the genes identified were known and represented characteristic hESC genes, such as POU5F1 and SOX2. However, others could not be matched to known Unigene clusters. The SAGE data also confirmed that biologically, mouse ESCs (mESCs) are essentially different to hESCs. Unlike mESCs, hESC expression of LIF and its receptor were very low, whereas POU5F1 and SOX2 expressionnearly 10 times higher. We also noted that hESCs also had elevated expression of proteins involved in the translation machinery, in the cytoskeleton, and in intercellular adhesion (CLDN6 and GJA1). In signaling, the wingless (Wnt), tumor growth factor-β and fibroblast growth factor pathways were active. We correlated the stemness characteristic to the expression of 21 putative hESC markers, which were either strongly expressed in hESCs or highly downregulated during differentiation. The successful identification of these genes will enable indepth studies of molecular pathways responsible for hESC characteristics such as pluripotency and repression of differentiation. However lots more work still needs to be done to provide more conclusive molecular markers for hESC controls. For example, in mice, expression of the zinc finger transcription factor REX1 is necessary for the maintenance of pluripotency; yet in one hESC line studied (HES4), REX1 was not identified during SAGE and quantitative polymerase chain reaction analysis, even though this line is capable of teratoma formation in immunodeficient mice and has been immortal in culture for >200 population doublings. One way to approach such a problem is to thoroughly document the incidence and times of activation or deactivation of candidate genes, using various profiling techniques, for example, massively parallel signature sequencing and microarrays. To date, the number of completed transcriptome investigations indicates that hESCs and their differentiated counterparts are highly uncharacterized. Such studies need to be pursued further to complete our understanding of the molecular regulation mechanisms within hESCs.
Undifferentiation and Differentiation The state of undifferentiation is one in which a hESC retains all of its characteristics, such as self-renewal and pluripotency. At the onset of differentiation, a hESC becomes increasingly restricted to developing down a definite pathway to form a specific cell type. Its telomerase expression is downregulated such that hESC lifespan becomes restricted as in somatic cells (86). Hence, the process of differentiation allows hESCs to acquire the properties of specialized cells and form almost all cells present in the human adult body. For example, the induced differentiation of hESC into functional cardiomyocytes has been documented by Mummery et al. (87), and in fact, hESCs favor spontaneous differentiation into tissues of either the neuronal or cardiomyocyte lineage. However, although it has been relatively simple to form hESC derivatives, the molecular regulation of differentiation is not well understood. Two approaches to hESC differentiation are available— directed and spontaneous. Spontaneous differentiation can occur if unfavorable culture conditions prevent the maintenance of undifferentiation, or if hESCs are cultured for extended
Stem Cell Reviews ♦ Volume 1, 2005
Human Blastocyst Culture and Derivation of ESC ___________________________________________________________________97 Table 3 Unanswered Questions With Respect to Clinical Application of hESC-Differentiated Cells How Should hESC-Differentiated Cells be Delivered? • Should the cells be injected at the site of injury? • Should the cells be injected into the portal vein to allow homing to the injured site? Is there an in vivo somatic stem cell niche? • How do niche cells influence stem cell fate decisions? What should be the cell dosage and what is its long term stability at the injured site? What are the side effects of hESC-differentiated cell therapy? • Will there be teratoma formation? • Should progenitor cells be delivered instead of terminally differentiated cells? • Could hESC- differentiated cells be purified by FACS and then delivered? How is in vivo function after delivery? How safe is hESC-differentiated therapy in terms of microbial transmission?
periods and/or at high density. Once hESCs lose their pluripotency, they eventually differentiate into tissues from the ectodermal, mesodermal, and endodermal germ layers. One of the outcomes of spontaneous differentiation is the formation of EBs—which are spheroid colonies composed of and which can express markers representing tissues from all three germ layers (88). Continuous EB culture results in the formation of lineages such as endothelium, skeletal muscle, and hematopoietic cells (89). Directed differentiation can be achieved by addition of differentiation agents (such as growth factors), by transfection with a gene that will regulate differentiation via a specific pathway, or by coculture with companion cells. For example, dopaminergic neurons can be produced when hESCs are cocultured with mouse stromal PA6 cells, and neuronal cells have been produced after treatment of hESCs with nerve growth factor and retinoic acid (90). Coculture is a useful approach as it is simple and fast, and the cells in direct physical contact with hESCs may express certain factors that induce differentiation into the required cell type. Transfection is a straightforward approach for isolating a pure and uniform culture of only the differentiated cell type, but can also be used to label cells for purification of the specific differentiated lineage (78,91). These methods can be carried out directly on the hESCs or on spontaneously differentiated EBs. One way to produce EBs is to enzymatically disaggregate hESC colonies and grow the resulting clumps in suspension (89), and then direct the uniform differentiation of a particular lineage by treatment with specific growth factors. For example, activin-A preferentially induces mesodermal expression (88). Subsequently, the required lineage can be recovered by cell sorting methods such as fluoresence activated cell sorting (FACS). The development of such systems will be essential for the specific and controlled differentiation of hESCs to generate cells and tissues for drug development, transplantation and gene delivery therapies (Table 3). Furthermore, it is hoped that by elucidating the mechanisms behind which cells differentiate into tissues, we will develop a better understanding of human developmental biology (92).
References 1. Edwards RG, Steptoe PC, Purdy JM. Br J Obstet Gynaecol 1980;87(9):737–756. 2. Bongso A, Ng SC, Ratnam S. Co-cultures: their relevance to assisted reproduction. Hum Reprod 1990;5(8):893–900. 3. Voluntary Licensing Authority. The second report of the Voluntary Licensing Authority For Human In Vitro Fertilization And Embryology, London, 1987. 4. Wong PC, Bongso TA, Ng SC, et al. J Obstet Gynaecol 1988; 19:41–43. 5. Yovich JL, Yovich JM, Edirisinghe WR. Fertil Steril 1988;49(5): 858–864. 6. Iritani A. Fertil Steril 1988;50:543–551. 7. Braude P, Bolton V, Moore S. Nature 1988;332(6163):459–461. 8. Menezo Y, Dale B. Hum Reprod 1995;10(6):1326–1328. 9. Janny L, Menezo YJ. Mol Reprod Dev 1996;45(1):31–37. 10. Bolton VN, Wren ME, Parsons JH. Fertil Steril 1991;55(4):830–832. 11. Bongso A, Ng SC, Fong CY, et al. Fertil Steril 1992;58(3):569–574. 12. Bongso A, Fong CY, Ng SC, Ratnam, S. Fertil Steril 1994; 61(5):976–978. 13. Bongso A, Fong CY, Ng SC, Ratnam S. Hum Reprod 1994; 9(11):2110–2117. 14. Bongso TA, Fong CY, Ng CY, Ratnam S. Cell Biol Int 1994; 18(12):1181–1189. 15. Bongso A, Fong CY. Curr Opin Obstet Gynecol 1993;5(5):585–593. 16. Liu HC, He ZY, Mele CA, Veeck LL, Davis O, Rosenwaks Z. Fertil Steril 1999;71(2):361–367. 17. Fukaya T, Chida S, Murakami T, Yajima A. J Exp Med 1996; 180(3):225–232. 18. Gardner DK, Lane M, Calderon I, Leeton J. Fertil Steril 1996; 65(2):349–353. 19. Phillips KP, Leveille MC, Claman P, Baltz JM. Hum Reprod 2000; 15(4):896–904. 20. Lane M, Hooper K, Gardner DK. J Assist Reprod Genet 2001; 18(9):519–525. 21. Bongso A, Fong CY, Ng SC, Ratnam S. Assist Reprod Rev 1995; 5:106–114. 22. Fischer B, Bavister BD. J Reprod Fertil 1993;99(2):673–679. 23. Maas DH, Storey BT, Mastroianni L. Jr. Fertil Steril 1977; 28(9):981–985. 24. Garris DR. Endocrinology 1984;114(4):1219–1224. 25. Hardy K, Handyside AH, Winston RML. Development 1989;107:597–604. 26. Muggleton-Harris AL, Findlay I, Whittingham DG. Hum Reprod 1990;5:217–220. 27. FitzGerald L, DiMattina M. Fertil Steril 1992;57(3):641–647. 28. Alikani M, Cohen J, Tomkin G, Garrisi GJ, Mack C, Scott RT. Fertil Steril 1999;71(5):836–842. 29. Garello C, Baker H, Rai J, et al. Hum Reprod 1999;14(10):2588–2595. 30. Sakkas D, Shoukir Y, Chardonnens D, Bianchi PG, Campana A. Hum Reprod 1998;13(1):182–187. 31. Salumets A, Hyden-Granskog C, Suikkari AM, Tiitinen A, Tuuri T. Hum Reprod 2001;16(10):2177–2181. 32. Tesarik J, Junca AM, Hazout A, et al. Hum Reprod 2000; 15(6):1396–1399. 33. Van Royen E, Mangelschots K, De Neubourg D, et al. Hum Reprod 1999;14(9):2345–2349. 34. Ziebe S, Petersen K, Lindenberg S, Andersen AG, Gabrielsen A, Andersen AN. Hum Reprod 1997;12(7):1545–1549. 35. Gardner DK, Vella P, Lane M. Fertil Steril 1998;69:84–88. 36. Gardner DK, Lane M, Stevens J, Schoolcraft WB. Fertil Steril 2001;76(6):1175–1780. 37. Jones GM, Trounson AO. Hum Reprod 1999;14(6):1405–1408. 38. Wittemer C, Bettahar-Lebugle K, Ohl J, Rongieres C, Nisand I, Gerlinger P. Hum Reprod 2000;15(12):2591–2597. 39. Bongso A, Fong CY, Matthew J, Ng LC, Kumar J, Ng SC. Asst Reprod 1998;9(2):70–78. 40. Ebner T, Moser M, Yaman C, Feichtinger O, Hartl J, Tews G. Fertil Steril 1999;72(4):599–603.
Stem Cell Reviews ♦ Volume 1, 2005
98 ________________________________________________________________________________________________Bongso and Tan 41. Van Blerkom J, Bell H, Weipz D. J Electron Microsc Tech 1990;16(4):298–323. 42. Scott L, Alvero R, Leondires M, Miller B. Hum Reprod 2000;15(11):2394–2403. 43. Smith AG. Annu Rev Cell Dev Biol 2001;17:435–462. 44. Fisch JD, Rodriguez H, Ross R, Overby G, Sher G. Hum Reprod 2001;16(9):1970–1975. 45. Isiklar A, Mercan R, Balaban B, Alatas C, Aksoy S, Urman B. J Reprod Med 2002;47(7):540–544. 46. Shapiro BS, Harris DC, Richter KS. Fertil Steril 2000;73(3):582–586. 47. Gardner DK. Hum Fertil 2000;3(4):229–237. 48. Fong CY, Bongso A. Hum Reprod 1998;14(3):774–781. 49. Testart J, Lassalle B, Belaisch-Allart J, Forman R, Frydman R. Lancet 1986;2(8):506–569. 50. Mohr LR, Trounson AO. Ann NY Acad Sci USA 1985;442:536–543. 51. Van Steirteghem AC, Van den Abbeel E, Camus M, et al. Hum Reprod 1987;2(7):593–598. 52. Bongso A, Fong CY, Ng SC, Ratnam S. Theriogenology 1984; 41:161. 53. Thomson JA, Itskovitz-Eldor J, Shapiro SS, et al. Science 1998;282(5391):1145–1147. 54. Pickering SJ, Braude PR, Patel M, et al. Reprod Biomed Online 2003;7(3):353–364. 55. Reubinoff BE, Pera MF, Fong CY, Trounson A, Bongso A. Nat Biotechnol 2000;18(4):399–404. 56. Heins N, Englund MC, Sjoblom C, et al. Stem Cells 2004;22(3): 367–376. 57. Suss-Toby E, Gerecht-Nir S, Amit M, Manor D, Itskovitz-Eldor J. Hum Reprod 2004;19(3):670–675. 58. Amit M, Itskovitz-Eldor J. J Anat 2002;200(Pt 3):225–232. 59. Richards M, Fong CY, Chan WK, Wong PC, Bongso A. Nat Biotechnol 2002;20(9):933–936. 60. Richards M, Tan S, Fong CY, Biswas A, Chan WK, Bongso A. Stem Cells 2003;21(5):546–556. 61. Hovatta O, Mikkola M, Gertow K, et al. Hum Reprod 2003;18(7): 1404–1409. 62. Mitalipova M, Calhoun J, Shin S, et al. Stem Cells 2003;21(5): 521–526. 63. Cowan CA, Klimanskaya I, McMahon J, et al. N Engl J Med 2004;350(13):1353–1356. 64. Li L, Baroja ML, Majumdar A, et al. Stem Cells 2004;22(4): 448–456. 65. Hwang WS, Ryu YJ, Park JH, et al. Science 2004;303(5664): 1669–1674. 66. Trounson AO. Fertil Dev 2001;13(7,8):523–532. 67. Daniels R, Hall V, Trounson AO. Biol Reprod 2000;63(4): 1034–1040.
68. Munne S, Grifo J, Cohen J, Weier HU. Am J Hum Genet 1994;55(1):150–159. 69. Bongso A, Richards M. Best Pract Res Clin Obstet Gynaecol 2004;18(6):827–842. 70. Xu C, Inokuma MS, Denham J, et al. Nat Biotechnol 2001; 19(10):971–974. 71. Amit M, Margulets V, Segev H, et al. J Biol Reprod 2003; 68(6):2150–2156. 72. Cheng L, Hammond H, Ye Z, Zhan X, Dravid G. Stem Cells 2003;21(2):131–142. 73. Fong CY, Santhanathan AH, Wong PC, Bongso A. Reprod Biomed Online 2004;9(3):321–325. 74. Tomlinson M, Sakkas D. Hum Reprod 2000;15:2460–2463. 75. Burdon DW. Issues in contamination and temperature variation in the cryopreservation of animal cells and tissues. Revco Technologies application note, 1999;99–108. 76. Clarke GN. Hum Reprod 1999;14(12):2941–2943. 77. Tedder RS, Zuckerman MA, Goldstone AH, et al. Lancet 1995;346(8968):137–140. 78. Reubinoff BE, Pera MF, Vajta G, Trounson AO. Hum Reprod 2001;16(10):2187–2194. 79. Richards M, Fong CY, Tan S, Chan WK, Bongso A. Stem Cells 2004;22(5):779–789. 80. Amit M, Carpenter MK, Inokuma MS, et al. Dev Biol 2000;227(2):271–278. 81. Draper JS, Smith K, Gokhale P, et al. Nat Biotechnol 2004; 22(1):53–54. 82. Pera MF. Nat Biotechnol 2004;22(1):42–43. 83. Buzzard JJ, Gough NM, Crook JM, Colman A. Nat Biotechnol 2004;22(4):381–382. 84. Inzunza J, Sahlen S, Holmberg K, et al. Mol Hum Reprod 2004;10:461–466. 85. Richards M, Tan SP, Tan JH, Chan WK, Bongso A. Stem Cells 2004;22(1):51–64. 86. Xu C, Jiang J, Sottile V, McWhir J, Lebkowski J, Carpenter MK. Stem Cells 2004;22(6):972–980. 87. Mummery C, Ward-van Oostwaard D, Doevendans P, et al. Circulation 2003;107(21):2733–2740. 88. Schuldiner M, Yanuka O, Itskovitz-Eldor J, Melton DA, Benvenisty N. Proc Natl Acad Sci USA 2000;97(21):11,307–11,312. 89. Itskovitz-Eldor J, Schuldiner M, Karsenti D, et al. Mol Med 2000;6(2):88–95. 90. Schuldiner M, Eiges R, Eden A. Brain Res 2001;913(2):201–205. 91. Muller M, Fleischmann BK, Selbert S, et al. FASEB J 2000; 14(15):2540–2548. 92. Bongso A. Handbook On Blastocyst Culture, 1st edition, Sydney Press Indusprint, Singapore.
Stem Cell Reviews ♦ Volume 1, 2005