ABSTRACT: The study of genomic imprinting in mammals started with analysis of parthenogenetic embryos. At the phenotypic level, embryos with two maternal genomes and no paternal genome proceed through early development unimpaired, and only begin to fail after implantation. The most recognizable early defect is reduced or non-existent trophoblast, the tissue that gives rise to the placenta. We applied the procedure for establishing Trophoblast Stem cells (TS cells) developed in the Rossant lab to parthenogenetic embryos, and were successful in making four different TS cell lines, three from MI oocyte derived blastocysts and one from MII derived blastocysts. Initial molecular characterization, including microarray analysis, indicates that these cells are indistinguishable from fertilized TS cells, with the single exception of null expression of the paternally expressed gene Snrpn. The only significant difference between parthenogenetic and fertilized TS cells was the frequency with which they could be derived. In our hands, fertilized blastocyst outgrowths produced TS cells at robust rates (10-12 colonies per blastocyst), while parthenogenetic blastocyst outgrowths produced only 4 colonies in 50 outgrowths, 100 times less frequently than fertilized embryos. This led us to hypothesize that those few TS cells that arose in parthenogenetic outgrowths were probably a result of very low frequency stochastic variation in the imprint status of a gene or genes required for either establishment or maintenance of stem cells, or both. The corollary to this hypothesis posits that the early failure of parthenogenetic embryos, and in particular parthenogenetic trophoblast, is a function of impaired stem cell function. This raises the intriguing possibility that microarray comparisons of parthenogenetic, fertilized and androgenetic blastocysts may reveal the identities of genes important for stem cell biology. To this end, my colleague Keith Latham, at the Fels Institute, Temple University, Philadelphia, made amplified cDNAs from pools of ten each of androgenetic, gynogenetic, or fertilized blastocysts. Three separate pools for each type of embryo were prepared for microarray analysis. The embryos were all produced by nuclear transfers between zygotes, a difficult technique that is Keith's special expertise. He used the Brady/Iscove protocol to generate quantitative 3' end biased cDNAs. We would like to compare the transcriptomes of these embryos using the MOE430 2.0 arrays. We expect that important insights into the biology of uniparental embryos in general, and stem cells in particular may be revealed. Keywords: other
