Constructing a fate map of the human embryo

The goal of our work is to understand how the initial stages of human development occur. Although the scientific community has learned a great deal about the factors that control the first developmental decisions in model organisms such as mice, very little is known about the parallel processes in humans. Much of what we have learned suggests that there are fundamental similarities and fundamental differences, but their extent has remained largely unexplored. We went to map this uncharted territory. How will we go about approaching this scientific question? We are studying the first few days of human development by growing embryos and studying their physical and molecular properties. Briefly, frozen human embryos, approximately the size of the head of a small pin, were donated to the UCSF Gamete and Embryo Bank by couples at the conclusion of fertility treatments. After written informed consent was obtained, we used these early embryos in our studies. When they are grown under the proper conditions, they continue to develop, expanding from two cells into aggregates that contain four, eight, sixteen, and eventually up to 100 cells without an increase in size. This process mimics what we think is happening during the first few days of an actual pregnancy. Thus, by studying the changes that occur during the first few days of human embryonic development in the laboratory, we think that we will gain a much better understanding of this process. During the current grant period, we made a great deal of progress toward accomplishing our overall goals. With regard to studying the physical properties of embryos, we showed that between the 8-cell and the 16-cell stage of development the component cells begin to segregate based on their ability to attach to one another. The cells on the outer surface of the embryo adhere tightly to one another. They also have a distinct orientation with morphological and molecular markers asymmetrically and systematically distributed at one end or the other. Once cells develop these specializations, we think that they are fated to form the placenta, which attaches the offspring to the uterus and supports its development for the rest of pregnancy. In contrast, cells on the inside of the embryo, which fail to develop these specializations, go on to form the so-called “inner cell mass” that develops into the offspring. In the coming year, we want to put these new findings into a molecular context. Specifically, we want to determine whether orientation of the cells that form the outer surface of the embryo happens before or after the cells start to express markers that suggest they have assumed a placental fate. These experiments will help us understand whether orientation or molecular signatures are the primary drivers of this initial developmental decision. We are also interested in the interplay of these two forces as related to the continued differentiation of the cells that go on to form components of the inner cell mass and, subsequently, all the cells of the offspring. What are the practical implications of this work? We think that stem cell researchers will use our findings to optimize protocols for differentiating human embryonic stem cells along the major lineages. Currently, we know that the most robust protocols for generating many types of cells (e.g., cardiomyocytes/muscle, neurons, pancreatic beta cells) involve triggering stepwise differentiation processes that recapitulate what happens during normal development. The roadmap that is generally used has been drawn using animal models. We think it will be very important to get equivalent information about the early stages of human embryonic development, which can be used to customize these protocols for the generation of differentiated human cell-based therapies. Our work also has relevance to treating human infertility. For example, it is very difficult to discern how embryonic development might be going awry when we know almost nothing about the analogous normal processes. Thus, we envision that the results of our work can be used as a backdrop for developing molecular correlates of embryo quality that can be used to identify the subset with the best potential for further development. We think that the ability to recognize and transfer these embryos will help improve pregnancy rates in couples who seek assisted reproductive technologies. Finally, establishing landmarks during the first few days of human development will help us optimize growth conditions for human embryos, an important consideration since the goal is to replicate as closely as possible the environment in which these initial developmental steps normally take place. In summary, we are making a great deal of progress in understanding the initial stages of human development. Thus far, our data suggest that the rules that have been learned studying animal models have been bent or reinvented in our species. Our continued work will further illuminate this principle.

Work from our group points to the interesting conclusion that differentiation during the early stages of human embryonic development may occur before any morphological specializations are evident. Our first clue that this might be happening came from a series of high definition, high magnification microscopic analyses in which we examined in detail the cells of five-day-old human embryos. The results were surprising. Specifically, the cells in the interior of the embryo that go on to form the entire body do not have an identical appearance. Instead, they seem to have unique specializations. For example the size of the nucleus that contains a cell’s DNA is variable as is its extent of DNA condensation, which is an estimate of its activity. These findings led us to suspect that cells of the early embryo might be unequal in their developmental potential, the primary theory that our project explores. During the current funding period, we made a great deal of progress toward testing this hypothesis. Our focus was on a set of 10 human embryonic stem cell lines that we derived from single cells of five sibling 8-cell human embryos. In two cases, multiple cells from the same embryo produced lines. Therefore, this collection consists of genetically identical or very genetically similar members. In one series of experiments, we profiled the genes that these lines express and compared them across the set. The results of this analysis suggested that some of the cells were at different stages of development when they were removed from the embryo for the purpose of line derivation. Interestingly, the differentially expressed genes included master regulators of development and important structural components that are linked to a cell’s identity. Our goal in the coming year is to determine if we can find early evidence of differentiation in human embryos, which would suggest that the changes that we are observing in cells that have been maintained in the laboratory are actually happening during normal early development. We were also interested in determining if our lines, which were derived from single cells that were removed from very early-stage human embryos, function differently than the majority of human embryonic stem cell lines that were produce from later-stage embryos. To address this question, we studied the expression of molecules that are associated with the first differentiation process that establishes the fate of early embryonic cells as either contributing to the body or the placenta. The latter transient organ connects the offspring to the mother and supports its development before birth. Initial experiments showed that the lines differed in their expression of molecules that are associated with establishing a placental fate. These included human chorionic gonadotropin, which is assayed to determine if a woman is pregnant. This finding suggested that some of the lines are more able to contribute to placental development as compared to formation of the offspring. To further investigate this possibility, we cultured lines that were derived from individual cells removed from a single embryo under conditions that stimulate placental development. We found that these human embryonic stem cells began to express markers that are normally associated with the first steps that establish placental identity. As differentiation continued, the cells began to express other placental factors that substantiated our theory that these lines were able to form both embryonic and placental structures. Thus, we think that deriving human embryonic stem cell lines from single cells that are removed during the very early stages of embryonic development yields lines that have an expanded developmental potential as compared to lines that are derived by conventional means from later-stage embryos. We think that our findings have interesting implications. At a basic science level, the suites of genes that the new lines differentially express as compared to other human embryonic stem cell lines could give us important clues about the factors that are active during the crucial initial stages of human development. We are very interested in testing the function of these molecules, which we theorize play important roles. Our data also add new information with regard to differences among the developmental potential of existing human embryonic stem cell lines. For example, they can be multi-potent with a relatively limited potential, pluripotent with the ability to differentiate into all the cells of the body, or totipotent with the ability to form both the placenta and the offspring. Our results suggest that human embryonic stem cell lines derived from very early stage embryos are closer to a totipotent state, which could make them more amenable to subsequent differentiation into many cell types, a theory we will be explore during the coming year.

During the past year, work in our group that was funded by this CIRM award has focused on three novel human stem cell systems. The first was a collection of 10 human embryonic stem cell lines that we derived from single cells, termed blastomeres, of very early stage human embryos. They were grown for three days in the lab by which time they were comprised of eight cells. Much of our work, described below, suggests that these cells have unique properties as compared to human embryonic stem cell lines that are derived by conventional means, i.e., from intact embryos that are grown for five–six days in the lab and are comprised of approximately one hundred cells. Of note, these lines were submitted to the NIH Registry in December of 2009. We were notified that they did not fit the Federal definition of a human embryonic stem cell line, which includes being derived from a five–six-day-old embryo. Therefore, a comment period was opened via the Federal Register regarding a proposed change to the definition of hESCs as coming from early embryos up to and including the 5-6 day stage. At this time the NIH is still considering the public comments. Therefore, the NIH cannot review for registration our new lines that were derived from single blastomeres and work employing this novel cell model is not eligible for Federal funding. Therefore, our experiments would not be possible without CIRM funding. Recently, we focused on comparing the global gene expression patterns of the lines. These “molecular fingerprints” give us important clues about differences in their potential with regard to forming the various cell types that are envisioned for use in regenerative medicine therapies. For example, some of the lines express very high levels of molecules that are only made by neurons. Others seem to be acquiring the characteristics of heart muscle or liver cells. We think that this finding is important because it should be easier to make particular differentiated cell types, such as those of the pancreas, from lines that are already predisposed to differentiate down this pathway. In additional experiments, we tested the theory that the individual molecular fingerprints of the blastomere-derived lines were a snapshot of differences that could be observed in human embryos at equivalent stages to those from which the lines were derived. Therefore, we detected molecules that were differentially expressed among the lines. In several cases, we observed expression in only a subset of the eight cells that comprise three-day-old embryos. This is evidence that, in humans, the developmental clock is running at a fast pace with differentiation evident at an earlier stage than was previously thought to occur. This finding suggests that later stage embryos from which nearly all the existing human embryonic stem cell lines have been derived may be comprised of cells that have further differentiated. The second novel human stem cell system was committed progenitors isolated from early gestation human placentas. We were very interested in pinpointing the location of the cells, which allowed us to devise procedures for purifying them. Then we used information that our group and other investigators have generated about signals that enable their self-renewal to devise conditions that support continuous growth of these progenitors in the laboratory. The new progenitor model will allow us to study important aspects of human placental development that were previously inaccessible. This work is important because the placenta plays a large role in governing pregnancy outcome. For example, this transient organ uses an unusual tumor-like invasive process to anchor the developing baby to the uterus. Additionally, the placenta transfers nutrients and wastes to and from maternal blood, respectively. Thus, with this new model we will be able to study the formation and the function of the cells that carry out these important tasks. The third novel human stem cell model emerged from our discovery that the lines described above, which were derived from very early-stage human embryos, could spontaneously form human placental stem cells, that is, the earliest stage precursors that ultimately give rise to committed placental progenitors. We used our knowledge of how to maintain the latter cells to devise conditions that enable growth of this population in the laboratory. Currently, we are refining methods that will enable their continuous self-renewal. We are also carrying out a detailed analysis of their developmental potential. Additionally, we want to establish banks of the cells for distribution to our colleagues who are also studying basic mechanisms of human placental development. Finally, we completed experiments in which we devised a novel method for growing human embryonic stem cells that does not require their exposure to other cell types, a potential source of infection. We discovered a molecule that could sustain them.

This is our final report. Our overall goal was to gain insights into the first few steps in human development, which begins with one cell and ends with formation of the entire body. This is a very complicated process that is difficult to understand if we only study the terminal steps. We theorized that by starting from the beginning, very close to the one-cell stage, we could reconstruct the molecular foundation on which the body plan is built. We think that this information is crucial for devising regenerative medicine therapies. We must be able to program human stem cells that are in the early stages of the developmental continuum into differentiated progeny that can integrate into fully formed organs and tissues for repair purposes. If we understand, at a molecular level, the processes that originally gave rise to the specialized cell types, then we can reproduce them in the laboratory, generating cells for transplantation into patients. We accomplished all the goals that were set forth in our original application. We proposed a systematic analysis of the early steps in human development. First, we used high-powered microscopy techniques to conduct a detailed examination of human embryos. We let them develop for 6 days in the laboratory. At this point the round embryo, which is about the size of a pinhead, consists of 2 types of cells. One type, which is found at the surface, is fated to form the placenta, which physically connects the offspring to the mother and regulates the transfer of substances to the offspring. The other type, a small cluster of cells at one end of the interior of the embryo, goes on to form the entire body. When we examined the components of this cluster at very high magnification we saw that the individual cells looked quite different from one another and some had mature characteristic. This was a surprise because similar examinations of embryos at the same stage from other species have shown that these clusters are comprised of cells that all look alike with quite primitive structures, suggesting that they are functionally very similar. Thus, we concluded that the initial stages of human embryonic development may be on a faster track than observed in animal models. How could we test this theory? Since we are working in the human species, there are a limited number of approaches available to us and we had to think of new ways to approach this question. Thus, we decided to take advantage of the fact that, working with colleagues, we helped to develop methods for deriving hESC lines from single cells of early-stage human embryos that were grown for only a few days in the laboratory. This method has many advantages as compared to the conventional approaches that have been used to establish nearly all the lines that scientists work with. For example, we know the precise cell that generated the line and its age (in days) at the time of derivation. This information is difficult to obtain when hESCs are derived from intact 6-day-old human embryos that we now think might be comprised of many kinds of cells. By making one change to this method, we made it very efficient. Specifically, we took into account the fact that the cells in the interior of the embryo reside in clusters. Therefore, when we cultured single cells that were removed from embryos we tried to recreate this environment by sandwiching them in molecules by which they are usually surrounded. Using this method, we extracted 8-12 cells from each of 5 embryos that had been grown in the laboratory for 3 days. Ten of these cells gave rise to hESC lines. One embryo produced 4 lines (UCSFB1-4), a second embryo gave 3 lines (UCSFB5-7), and 3 embryos gave one line each (UCSFB8, UCSFB9 and UCSFB10). The 5 embryos were donated by one couple. Thus, the lines were either genetically related as are siblings or identical, i.e., triplets and quadruplets. The lines were registered with CIRM and are pending approval by NIH. What did we learn from this collection of cells? First, we compared their gene expression patterns, a bar code consisting of about 25,000 elements that is unique to every cell type. We found that these cells had bar codes that were different from hESC lines that were derived by conventional methods. Very interestingly, each cell line had a different bar code even if they were established from cells that were removed from the same embryo. This suggested that the lines might have retained differences that were present from the beginning. To test this theory, we looked for the same differences in cell clusters of early-stage human embryos and found them. This is additional evidence in support of our fast track theory of human embryonic development and gives us insights into the molecules involved, i.e., the differentially expressed portions of the bar code. We think that this information will be very useful for guiding the production of specific cell types that will be used in regenerative medicine therapies.