6

Exploring Opportunities and Challenges with Mammalian Embryo Model Systems

The final session of the workshop featured a closing keynote lecture from Martin Pera and a panel discussion focused on future opportunities and challenges with mammalian embryo model systems. Panelists were asked to summarize the lessons learned and the topics discussed throughout the workshop. They also considered novel strategies that could be moved forward to help advance the field and the level of fidelity that exists between model systems of human embryos and bona fide human embryos. The session was moderated by Nicolas Rivron. The final panel discussion included Ali Brivanlou, Jianping Fu, Kathy Niakan, and Mana Parast. Janet Rossant delivered closing remarks to wrap up the workshop.

HUMAN PLURIPOTENT STEM CELLS, THE HUMAN EMBRYO, AND THE SELF-RENEWING STATE

In his closing keynote lecture, M. Pera explored the developmental status of human pluripotent cells. Two fundamental questions about embryo models that remain unanswered were highlighted: the developmental status of human pluripotent stem cells and the stage of mammalian development that these cells most closely resemble. Understanding pluripotent cells in a developmental context is important for (1) comparative studies of early development, (2) the use of those cells as models of human embryonic development, and (3) understanding how best to direct lineage specification and differentiation, M. Pera said. Furthermore, it is not clear whether work on developing embryo models is starting with cells at the optimal

pluripotent state or whether these cells are epigenetically or otherwise off-center.

The current understanding of the various states of cell pluripotency is largely based on work in the mouse, M. Pera said. Mouse naïve cells are embryonic stem cells (ESCs) maintained under conditions that strongly suppress differentiation (E4.5 epiblast). Epiblast stem cells (EpiSCs) correspond to late primitive streak stage (E7.5), which is almost the last stage in which pluripotent cells are seen in the embryo. Epiblast-like stem cells are an unstable cell type that corresponds to formative pluripotency in the early post-implantation epiblast (E5.5). These are naïve cells that have gone through a transition that then enables them to respond to the signals that will drive lineage specification. The naïve state, or ground state, is a stage in embryonic development at which the pluripotent cells are capable of forming all somatic tissues and the germ line and are lacking in any bias toward differentiation into any particular fate in the pre-implantation epiblast. This state can be maintained in vitro by strong pharmacologic blockade of signaling pathways that drive differentiation. In a mouse, only naïve pluripotent cells are capable of generating germ line chimeras. Mouse EpiSCs correspond to the late gastrula-stage embryo and show lineage priming (i.e., co-expression of pluripotency genes along the lineage-specific genes), M. Pera said (Kojima et al., 2014).

There have been some attempts to create human naïve cells, M. Pera said. Human ESCs and induced pluripotent stem cells (iPSCs) resemble in some respects mouse EpiSCs, but not mouse ESCs, and human ESCs and iPSCs are widely considered to be equivalent to them. However, this interpretation is based on relatively limited interspecies comparative data and does not take into account heterogeneity in stem cell populations. Human naïve PSCs equivalent to mouse naïve cells have not yet been described, though some cell lines similar to human pre-implantation epiblast have been developed. The maintenance of genetically stable human naïve PSCs has proven problematic, and mouse naïve cells may be epigenetically unstable under some conditions. The embryonic equivalent of the stable attractor state represented by human lines and conventional lines remains unknown, M. Pera said. He calls this a “stable attractor” state because regardless of whether the lines are made from the human teratocarcinoma, an embryo, or from reprogramming, under a wide range of culture conditions they revert and settle into this particular state, which is the conventional human stem cell. Assessment of the developmental status of human pluripotent cells is complicated by heterogeneity in cultures, as it was originally in the mouse, and there are subpopulations of pluripotent cells within the human cultures with distinct biological properties and gene expression (Kolle et al., 2009; Laslett et al., 2007). Only a minority of the cells in human PSC cultures show a high capacity for self-renewal.

Human pluripotent stem cells grown in the conventional fashion that have a high capacity for self-renewal also have a metabolic and cell cycle status similar to the mouse peri-implantation epiblast, M. Pera said. Transcriptionally, conventional hPSCs with high self-renewal capacity correspond most nearly to peri-implantation primate epiblast (E13–E16). In terms of development capacity, conventional hPSCs with high self-renewal capcity are competent to give rise to germ line cells similar to early post-implantation mouse epiblast formative cells. Self-renewing hPSCs do not resemble mouse epiblast stem cells and are not primed for any particular fate, he added. When the embryo implants in the mouse, the epiblast cells are no longer capable of differentiating to extraembryonic trophoblast or primitive endoderm. Whether the same developmental restriction occurs in the primate is unclear. Several other questions that remain to be addressed include will clonal growth and genetic stability improve if it is possible to maintain self-renewing subpopulations of hPSCs in pure form in culture. Much of the variation in the differentiation capacity of iPSC lines is genetically determined by the background of the donor. However, it is yet unknown if the variation in the differentiation capacity of hPSC lines arises from variation in the stability of substrates within the cultures.

FINAL PANEL DISCUSSION

Fidelity of Model Systems

Stem cells have been used to form different embryo models that recapitulate specific windows of development, Rivron said. However, these models are still crude in that they only recapitulate certain features of development rather than capturing the entire process. He asked the panelists to comment about the fidelity of the models presented (i.e., the extent to which they represent what they are intended to model) and about how those models could be made more reliable. For models of the earliest stages of development, questions of fidelity are among the most challenging to answer, Brivanlou said. Assessing fidelity for his models would require specimens that are samples of normal human development post-attachment, which are not accessible. The best option is to morphologically compare a given state of embryogenesis with the limited number of pictures that are available from the Carnegie or Kyoto embryo collections.¹ Another barrier to evaluating the fidelity of his models, Brivanlou said, is the lack

___________________

¹ More information about the Carnegie Institution of Washington’s collection of human embryos can be found at https://embryo.asu.edu/pages/carnegie-institution-washington-department-embryology (accessed April 2, 2020), and for a more detailed history of the Kyoto Collection of Human Embryos and Fetuses, see Yamaguchi and Yamada (2018).

of control over the genetic backgrounds of his specimens or other variables based on ethnicity, diet, and metabolism, for example. Rivron observed that benchmarks are needed to understand what happens in real human embryos so those benchmarks can be used to verify these models. Studying human embryos is also important in order to examine the genetic variability that naturally occurs in humans but is poorly recapitulated in stem cells, Rivron added.

Advantages of Modeling Across Species

There is room for improvement in prolonging in vitro culture of human and nonhuman primate embryos, Fu said. He suggested that, in principle, nonhuman primate embryos could be cultured to establish benchmarks for calibrating in vitro human stem cell models. Golos suggested analyzing archived material from nonhuman primates for markers that are being discovered in extended culture or pluripotent cell studies. He also suggested that extended culture of nonhuman primate embryos has great potential to complement human studies, and he noted that in the past, human advances such as in vitro fertilization have been translated to the nonhuman primate. There are approaches in human embryology that can be complemented by applying human advances to other species, Brivanlou said, and creating this type of feedback loop may help to further our understanding of human embryology. This will require improving the comparative ability to interpret human development based on what is thought to occur in other systems, he said. Furthermore, a limitation that transcends the human system and affects all embryogenesis is the existence of critical dimensions and variables that are being missed in current measurements. For example, Brivanlou said, simple factors heretofore unconsidered (e.g., contour, mechanical forces, tissue stiffness) are now emerging as major instructive factors in guiding cell fate, path information, and morphogenesis. Model systems are useful in uncovering these types of new dimensions in other species and applying them to study human origins in particular, and embryology more broadly. It would be a mistake to consider all nonhuman primates together, Parast said. Many species have yet to be studied, and it is likely that some will be better than others in recapitulating different elements, such as placentation or trophoblast differentiation of the pluripotent stem cell, she said.

Open Questions with Embryonic Development

Despite progress in understanding development in general, large mysteries remain, Renee Reijo Pera said. Brivanlou’s philosophical perspective is that everything is a cycle within a cycle: the union of the sperm and egg gives

rise to an adult who produces sperm and eggs that give rise to another adult. Similarly, generating a model of human embryos began by using ESCs, which themselves came from an inner cell mass of a blastocyst. In that sense, it is almost like sending the human ESCs “back home” when they are put into the three-dimensional (3D) context of an embryo. “The fact that this very strong aspect of self-assembly and self-organization seems to be the driving force behind that, is not only mysterious, but also really beautiful,” Brivanlou said, “and there is something in there that is beyond our scientific measurements and evaluations.”

Fidelity of Three-Dimensional Systems

M. Pera asked the panelists about the extent to which 3D systems—which are presumed to be closer to normal embryonic conditions—are actually better in terms of fidelity or function than previous, cruder approaches. This gets close to the contentious question of whether researchers make models that are better than the original, Brivanlou said. Creating a model that is better than reality could be advantageous, for example, if it fixed something that is aberrant (e.g., curing a disease) in a way that is not otherwise possible. The issue is much more complicated in borderline cases, he said. Brivanlou went on to question the extent to which human or primate embryological work can be rationalized or justified based on utilitarian grounds versus the desire for basic knowledge. In his experience, he said, support and funding are not generally forthcoming in response to arguments based on the need to understand our human origins; utilitarian arguments based on the potential to cure disease, for example, tend to be more effective.

Further investment is needed to characterize the basics of early development across many species, including humans, Niakan said. Fundamental mechanisms of the earliest stages of human development are only beginning to be elucidated, but knowledge about the extent to which these mechanisms are conserved across species is needed before they can be recapitulated in vitro, she said. That knowledge will also be required to make improvements to directed-differentiation protocols that will be needed to make cells of potential therapeutic relevance and start to accurately model diseases. In many directed-differentiation protocols, the cell types produced are early precursor progenitor cells and not the mature cells that are needed, she said. Her philosophy is that “one never knows what will ultimately arise from all of that knowledge, but the knowledge itself is extremely important.”

Contribution of Model Systems

The epigenetic status of human systems (e.g., variable methylation) is a complicating factor, M. Pera said, and it might be rectified by culturing cells in more natural conditions. Rather than trying to identify or generate better cells, Fu said his group is trying to develop model systems that allow for obtaining embryonic antigen-like cells in 3D culture systems. That knowledge could then be used to understand how, in the human context, embryonic ectoderm drives epiblast gastrulation and primordial germ cells develop in the embryonic ectoderm compartment. Synthetic models provide experimental systems for studying those fundamental biological questions, which was not possible in the past, Fu said. Further novel and important uses for these systems are to study human placentation, miscarriage, and later complications of pregnancy, such as fetal growth restriction and preeclampsia, Parast added. She noted that this area of research has been underfunded and overlooked—in part because it is considered controversial. However, she suggested that framing this research in the context of improving pregnancy outcomes and improving life for the next generation could help demonstrate the importance of this type of work.

USING EMBRYO MODELS TO UNDERSTAND HUMAN DEVELOPMENT

In addition to understanding infertility, early embryo loss, and birth defects, M. Pera suggested that investigating the embryonic origins of adult diseases is another important rationale for studying the human embryo. This is a controversial area of research, he said, but evidence suggests that epigenetic programming during early development affects the onset of disease in later life, particularly cardiometabolic and neurological disorders. Embryo systems could be used to address these issues, he suggested, as well as to provide opportunities to improve the fidelity of stem cell differentiation. Two decades ago, when his laboratory and others first started to derive cell lines from the human blastocyst, they were asked to justify the use of human embryos in this research, M. Pera said. In Australia, the question was not whether researchers would be funded by the government to do this research; the question was whether they would be put in jail for doing it, he said. One of the arguments researchers there made was that embryo systems would ultimately be useful for informing the understanding of human development, and based on what he heard during the workshop, he said, the prospects for achieving that have never been better.

In her final remarks, Rossant underscored the wealth of opportunities that are now available to study human development in new ways using

actual human embryos and stem cell models in culture systems. However, progress remains to be made to understand enough about the early human embryo to establish benchmarks and comparators. Regarding stem cells themselves, much remains to be understood about the state of those cells and how to maintain those states in the desired way. “We are not making human embryos in a dish; we are not making test tube babies; we are not generating anything that could in any way be considered as viable human embryos, nor is that the intent of the any of the research that you heard about today,” Rossant emphasized. Instead, the aim is to study the important phases of human development that are relevant to issues of human infertility, placental problems, and disease. Embryo model systems provide a tremendous opportunity, she said, to better understand human development and to develop new tools to study disease.

This page intentionally left blank.