Human embryonic stem cells remain the focus of an ever-intensifying public debate that blurs the limits of biology, confusing cultured tissues with children, and blastocysts descended from fertilized ova with those derived from somatic cell nuclei. When Woo Suk Hwang's group from Seoul National University published controversial findings two months ago, they used ultraspecific language.1 The 11 stem cell lines were created not from cloned human embryos, but "nuclear transfer constructs." Nevertheless, for a field that demands succinct descriptions, human embryonic stem cells (hESCs) themselves still defy definition. Researchers continue to tweak decades-old culture techniques to derive and better characterize the potentially powerful self-renewing populations of cells. But a gap remains between what is known about the cells and what they actually are.
Do they have a counterpart in the embryo? And if so, from what developmental line-age do they descend? Or are they more an artifact of their artificial environs, the culture dish? Recent gene-expression profiling of hESCs offers glimpses of sets of activated genes, and mostly complements early descriptive embryology of the mouse. But the portrait of the hESC is still a fuzzy one.
Human ESCs are nurtured in vitro from the inner cell mass (ICM) cells of 5-day-old blastocysts. But how the finished product relates to those ICM precursors "is a big mystery," says Renee A. Reijo Pera, codirector of the program in human embryonic stem cell biology at the University of California, San Francisco, who favors the embryonic counterpart hypothesis. Still, "we know that there is no cell in the embryo that is identical to an ES cell," says John Gearhart, codirector of Johns Hopkins' Institute for Cell Engineering. "The issue then is, 'What is the cell that gives rise to an embryonic stem cell, and how close can we get to it?"' Although research into clinical applications of hESC descendants is already well under way, knowing more about the basic nature of the beginning cells might speed the development of regenerative medicine.
Describing a hESC is like identifying a person. Who an individual is can be seen in where he or she came from; physical characteristics; and his or her role in the community. ESCs can likewise be defined in terms of their origin, form, and function in a tissue.
Pluripotent cells are uncommon and ephemeral in embryos, nestled among blastomeres, ICM cells, primitive ectoderm, and primordial germ cells (PGCs). They are essential to maintaining the high cell-division rates that build a new body. The culturing process isolates, captures, and sustains these abilities. So to understand ESCs, a good start is to trace embryogenesis to see which cells they most closely resemble. And so far, it looks like hESCs are not most like ICM cells.
An early clue that hESCs might most closely resemble PGCs, or their precursors, came from Leroy Stevens, who worked on teratoma-prone mice at the Jackson Laboratory from the 1950s through the 1980s.2 Stevens meticulously dissected mouse embryos in search of the cells that led to formation of teratomas – hideous multitissue growths, usually in a testis. Jackson's critical leap was recognizing the close resemblance of the responsible pre-PGCs to cells of the early embryo. In a masterful series of experiments, he implanted early embryo cells, including those of the ICM, into adult testes. As he anticipated, teratomas formed. Stevens named these ICM cells "pluripotent embryonic stem cells," hence coining the now-familiar term, but it was quickly forgotten because the cells also gave rise to cancer cells in some teratomas. The besmirched cells became known as embryonal carcinoma, or EC, cells. A teratoma without EC cells is benign; with EC cells, it is a teratocarcinoma. EC cells proved difficult to nurture in blastocysts to see if they could support development because of chromosomal instability. Nevertheless, they could become part of normal chimeras on rare occasion.
A Unique Set of Markers is Expressed in Human ESCs Including:
• Alkaline phosphatase
• Deleted in azoospermia-like (DAZL)
• Enhancer of rudimentary homolog (ERH)
• High mobility group AT-hook 1 (HMGA1)
• Octamer-binding protein (OCT4)
• POU domain class 5, transcription factor 1 (POU5F1)
• SRY box-containing gene 2 (Sox-2)
• Stage-specific embryonic antigens (SSEA-3, SSEA-4)
• Trafalgar (aka tumor-rejection antigen) (TRA-1-60, TRA-1-81)
In 1981, Gail R. Martin at UCSF, and Martin Evans and Matthew Kaufman at Cambridge University, isolated EC cells that had stable chromosomes and had apparently escaped the cancer taint.34 These were mouse ESCs, which became the foundation of knockout technology and led the way to the derivation of hESCs.56 Amander T. Clark, a research geneticist in Reijo Pera's group, distinguishes between these two types of pluripotent cells. "EC cells have chromosome abnormalities, and this is probably why you run into trouble when injecting them into blastocysts. ESCs do not have the chromosomal abnormalities, but have the same potential as EC cells."
This past year, Reijo Pera, Clark and their colleagues compared transcriptional profiles of three hESC lines to human ICM cells, painting a molecular portrait that supports the germ-cell origin hypothesis.7 "Markers of germ cells are expressed in hESCs and not in ICM cells. The implications are that hESCs are likely derived by selection of a retained small population from the ICM, or by ICM cells developing beyond their original stage," says Reijo Pera.
The researchers discovered, for example, that undifferentiated hESCs produce DAZL, which is a germ-cell specific marker that acts premeiotically. They also found that hESCs implanted in mouse peritoneal cavities, where they give rise to embryoid bodies, display more differentiated germ-cell markers, such as VASA. Moreover, while hESCs and ICM cells express three markers in common (NANOS1, STELLAR, and OCT4), hESCs alone sport certain stage-specific embryonic antigens. And so assuming that hESCs are most like ICM cells because that is their immediate source may be more guilt-by-association than reality.
Even if hESCs most closely resemble PGCs, or pre-PGCs, they are not the same as these in vivo cells. ESCs self-renew longer than PGCs, and, when placed in chimeras, support normal development. Culture conditions probably help to mold stemness.8 Clarifying the relationship between ESCs and PGCs, however, requires observing development of human embryos into what some consider unethical territory – beyond 14 days, when formation of the primitive streak heralds the first inklings of a nervous system.
Beyond the two most important defining characteristics-self-renewal and pluripotency-stem cells also have multiple nucleoli, a low cytoplasm-to-nucleus ratio, and lipid bodies in the cytoplasm. In a general sense, hESCs express 20% to 30% of the genome, compared to 10% to 20% for most somatic cells. The most highly expressed genes in hESCs encode proteins involved in cell-cycle control, apoptosis, DNA repair, development, and stress responses – a rather generic laundry list. But by using DNA microarrays, serial analysis of gene expression (SAGE), and the more powerful massively parallel signature sequencing (MPSS), researchers are homing in on specific genes whose products impart a state of stemness While DNA microarrays can only glimpse the sequences chosen for probing, SAGE and MPSS pull out thousands to millions of transcripts from a cell, and can spot splice variants.
In one recent study, Ariff Bongso, research professor in the department of obstetrics and gynecology at the National University of Singapore, and colleagues used SAGE to compare two hESC lines and a mouse ESC line. They identified 21 candidate hESC marker genes, including 10 novel ones. The two human lines were very alike in gene expression, despite their derivation from individuals of different sex and ethnicity.9
Whatever hES cells are, or aren't, might become clearer in mid-August, when researchers from 17 labs representing 10 countries present profiles of 75 hESC lines at the first meeting of the International Stem Cell Initiative at the Jackson Laboratory in Bar Harbor, Maine. Of special interest, says coordinator Peter Andrews, professor of biomedical science at the University of Sheffield, is the diversity of hESCs compared to the few closely related mouse strains.
Until then, actual description of hESCs remains murky. Says Hayek, "Human ESCs are real. What is artificial is their environment. And that is what will prove an extremely difficult nut to crack."
A Cultured Existence
Human embryonic stem cells need help in staying undifferentiated. Life on a dish typically requires some combination of mouse embryonic fibroblasts, extracellular matrix (ECM) components, antibiotics, amino acids, β-mercaptoethanol, leukemia-inhibiting factor, and growth factors. But rodent material may introduce pathogens, as well as sialic acid antigens, against which many people make antibodies.
Just why human embryonic stem cells (hESCs) need this support remains unclear. "Nobody knows for sure, but ECM components play a very important role not only in attachment but also for maintaining pluripotency and proliferation," says Miodrag Stojkovic, a professor in the Center for Stem Cell Biology and Developmental Genetics at the University of Newcastle upon Tyne. His group uses fibroblast-like cells from spontaneously differentiated hESCs as feeders.1 Other variations on the theme include new hESC lines exposed to murine ECM extracts, but not whole cells,2 and human foreskin fibroblasts.3 The Korean group that caused a stir two months ago when they created several hESC lines from cloned human blastocysts used human feeders both from nuclei donors and unrelated subjects.4 Nevertheless, the authors note that animal materials were used in some culturing conditions.
The need for feeders suggests an experimental route to defining stemness that complements transcriptional profiling: Remove supportive components one-by-one and see what happens. When researchers at Geron Corp., in Menlo Park, Calif., removed basic fibroblast growth factor (bFGF), within six weeks the hESCs began to differentiate – implicating bFGF in maintaining pluripotency.5
For Alberto Hayek, professor of pediatrics and director of the islet research laboratory at the University of California, San Diego's Whittier Institute, identifying activin-A as key to hESC maintenance was more serendipitous.6"We had been doing experiments to induce differentiation of human fetal pancreas β cells from human ESCs. We added the growth factors we knew could induce differentiation and found, much to our surprise, that the cells remained undifferentiated," he says. After eliminating all known markers of pluripotency, individually and in pairs, "we ended up with activin A." Take it away, and the cells differentiate.
Keeping hESCs undifferentiated also opposes the natural tendency for dividing cells to mutate and be selected for. In extended culture, hESCs tend to gain an extra chromosome 12 and short arm of chromosome 17 – the same glitches seen in embryonal carcinoma (EC) cells (see main article).7 That isn't necessarily alarming, says Peter Andrews, professor of biomedical science at the University of Sheffield, but may reflect the pluripotentiality of ESCs and EC cells. "If a mutation increases self-renewal capacity instead of the decision to drop dead or differentiate, there will be a selective advantage. Over time, cells with that change dominate the culture. This could happen in an EC cell in a tumor as well as in an embryonic stem cell."
Culture conditions could select different changes than those that would occur in vivo, but that's not alarming either, Andrews says. Consider hESCs in which the
- Ricki Lewis