|Reconstitution of the Mouse Germ Cell Specification Pathway in Culture by Pluripotent Stem Cells|
Protocols for the differentiation of pluripotent stem cells to functional therapeutically relevant cells often give relatively low yield. However, new protocols from various groups have addressed this by going back to learn more about the in vivo development of the cell type or tissue that they wish to attain, and utilising this knowledge in new more advanced differentiation protocols which aim to increase yield. In vitro attempts to generate gametes or primordial germ cells in mouse and human (Area reviews in Daley and Saitou et al) have largely been based on the isolation of these cells from spontaneously differentiating embryoid body cultures, an inefficient method that generally does not derive sufficient cells for substantial analysis. Now, researchers from the group of Mitinori Saitou at Kyoto University, Japan have demonstrated the efficient generation of PGC-like cells in mice which have spermatogenic capability. By utilising what information is known regarding temporal and signalling dynamics during PGC development in vivo and in vitro, they have devised a new protocol which leads to the development of PGC-like cells over multiple stages which reflects the development of the epiblast in vivo. The major novel step is the conversion of mouse embryonic stem cells (mESCs) to epiblast-like cells (EpiLCs), a state which is highly similar to cells of the pregastrulating epiblast, but distinct from epiblast stem cells (EpiSCs), which are known to be competent to express Blimp1/Prdm1 and Prdm14 (Kurimoto et al, Ohinita et al, Vincent et al and Yamaji et al), two of the major regulators of PGCs. Importantly, these regulators mediate the generation of functional sperm after ex vivo induction by BMP4 and neonatal intratesticular transplantation (Saitou et al). This study is published in the August edition of Cell (Hayashi et al).
Initial reasoning deemed that ground state (naïve) ESCs (as compared to the primed pluripotent state of EpiSCs) (Nichols and Smith) might rapidly differentiate into pregastrulating epiblast-like cells with the ability to undergo differentiation to a PGC fate using conditions similar to those used to induce EpiSC-like cells (Nichols et al, Guo et al, Han et al and Ying et al). Therefore, ESCs derived from embryonic day 3.5 blastocysts bearing Blimp1-mVenus (BV) and Stella-ECFP (SC) transgenes were stimulated with Activin A, bFGF, and 1% knockout serum replacement (KSR) resulting in the uniform induction of ESCs into flattened epithelial structures resembling the epiblast. These induced epiblast-like cells (EpiLCs) did not express BV or SC over the three day treatment period but immunofluorescence and quantitative-PCR analysis indicated that EpiLCs demonstrated properties that are consistent with pregastrulating epiblasts. It was next examined whether these EpiLCs could be induced into PGC-like cells under conditions that can induce epiblast cells to differentiate into the PGC fate (floating culture in GMEM with 15% KSR with cytokines including BMP4) and it was found that 2 day EpiLCs were highly competent to express Blimp1 in response to BMP4 and for subsequent healthy growth, suggestive of PGC induction as observed in vivo. Upon extended culture with BMP4 day 2 EpiLCs were induced into BV+SC+ PGC-like cells, with LIF enhancing the maintenance/proliferation of the BV+SC+ cells and more robustly by the combinatorial effects of LIF, SCF, BMP8b, and EGF. These PGC-like cells (PGCLCs) appeared similar, if not identical, to epiblast cells with regards to structural development and their competence to express Blimp1 in response to BMP4. Further, gene expression dynamics associated with PGCLC induction are very similar to those associated with PGC specification. Global transcriptional profiling of amplified RNA followed by unsupervised hierarchical cluster analysis showed that PGCs and PGCLCs clustered very tightly, indicating that PGCLC formation from ESCs through EpiLCs is a recapitulation of PGC formation from epiblasts. Epiblast, EpiLC and ESC also clustered together, and were similar to the PGCs and PGCLC when compared to EpiSC which clustered away from these sets of cells indicating their divergence from the other cell types. Further gene expression analysis showed that EpiSCs upregulated more genes associated with the development of a variety of organ systems (heart, blood vessels, kidneys, muscle, and bone) than E5.75 epiblasts and 2 day EpiLCs, demonstrating that EpiSCs acquire more developmentally advanced characteristics than E5.75 epiblasts and 2 day EpiLCs.
Immunofluorescence analysis of repressive epigenetic modifications of histones and DNA revealed that PGCLCs at day 6 appeared to have reduced H3K9me2 and cytosine methylation (5mC) and elevated H3K27me3 levels. In ESC to EpiLC differentiation, the H3K9me2 and 5mC levels increased and H3K27me3 level decreased, while in EpiLC to PGCLC differentiation H3K9me2 and 5mC levels decreased significantly and H3K27me3 levels increased, similar to that observed during PGC formation. Imprinting analysis demonstrated that a few sequences showed decreased levels of methylation suggesting that PGCLCs may be initiating the process of imprint erasure, as the global decrease of 5mC with a relative maintenance of imprinting in PGCLCs is a characteristic that is consistent with that of migrating PGCs. Analysis of PGCLC induction and proliferation indicated that BV induction from EpiLCs is an efficient process and, upon SC initiation, the cells proliferate slowly, a key characteristic of migrating PGCs alongside arrest at the G2 phase of the cell cycle. This provides further evidence that PGCLCs bear equivalent properties to PGCs and provides further evidence that PGCLC formation is a reconstitution of PGC formation.
The most exciting part of the study was the assessment of the ability of PGCLCs to undergo spermatogenesis, examined by their transplantation into the seminiferous tubules of W/Wv neonatal mice lacking endogenous germ cells. While non-sorted cells led to teratoma formation, three of 6 mice transplanted with BV+ cells harbored seminiferous tubules showing complete spermatogenesis, with the efficiency of PGCLC colonization comparable to that of PGCs in vivo. Oocytes were then fertilized with PGCLC-derived sperm, leading to normal zygote development which, by the blastocyst stage, exhibited strong expression of SC derived from the donor genome. Transfer of the embryos to foster mothers allowed the development of grossly healthy offspring with normal placentas and imprinting patterns, with both male and female offspring growing into normal fertile adults. The BV and SC transgenes were positive in 13 and 7 of 21 offspring, respectively, consistent with the transmission of the transgenes through haploid donor spermatozoa.
Cell surface markers (SSEA1, PECAM1, EPCAM, N-cadherin, Integrin-b3, Integrin-aV, CXCR4, and KIT) and their combinations were then screened to allow for the identification of those that define the BV+ population, to allow for the removal of transgenes from the initial ESCs as a selection method. SSEA1 and Integrin-b3 were found to be able to purify PGCLCs with essentially no contamination of teratogenic cells, establishing the formation and purification of PGCLCs from ESCs without relevant transgenic markers. These markers were then used in the analysis of PGC development from three iPSC lines (20D17, 178B-5 and 492B-4) all bearing Nanog-EGFP (NG) transgenes. Using the same differentiation protocol as previously used with ESCs followed by sorting with SSEA1 and Integrin-b3, the 20D17 line exhibited a prominent gene expression correlation with BV+ PGCLCs, while the other two lines showed a lower correlation. Only the 20D17 line exhibited proper spermatogenesis (3 of 18), and with no teratomas in the recipients of this line. The resultant sperm contributed to fertile offspring although some of the offspring died prematurely, apparently due to tumors around the neck region. Taken together, these findings demonstrate that, although iPSCs exhibit different induction properties depending on the lines, they can nonetheless form PGCLCs with PGC-like function.
Overall, this paper represents a huge step in the right direction for gamete differentiation from pluripotent cells. It eases the problem of low cell numbers for various analyses, including the identification of mechanisms behind PGC specification/proliferation/survival, the mechanism of epigenetic reprogramming in PGCs and the key proteins involved. Whether these results can be repeated with human ESC is the next big question. The authors note that differences between hESCs, mESC and mEpiSCs may hinder this move, and they suggest that the use of primate models may play a critical role. One caveat was the moderately disappointing results while attempting to repeat these studies in iPSC lines. This again suggests that iPSCs may not, at this stage and with current technology, be an appropriate replacement for ESCs in a therapeutic context for the scope of this study, if this data can indeed be moved into human models.
Guo G, Yang J, Nichols J, Hall JS, Eyres I, Mansfield W, Smith A.
Hayashi K, Ohta H, Kurimoto K, Aramaki S, Saitou M.
Han DW, Tapia N, Joo JY, Greber B, Araúzo-Bravo MJ, Bernemann C, Ko K, Wu G, Stehling M, Do JT, Schöler HR.
Kurimoto K, Yabuta Y, Ohinata Y, Shigeta M, Yamanaka K, Saitou M.
Nichols J, Silva J, Roode M, Smith A.
Nichols J, Smith A.
Ohinata Y, Payer B, O'Carroll D, Ancelin K, Ono Y, Sano M, Barton SC, Obukhanych T, Nussenzweig M, Tarakhovsky A, Saitou M, Surani MA.
Saitou M, Barton SC, Surani MA.
Vincent SD, Dunn NR, Sciammas R, Shapiro-Shalef M, Davis MM, Calame K, Bikoff EK, Robertson EJ.
Yamaji M, Seki Y, Kurimoto K, Yabuta Y, Yuasa M, Shigeta M, Yamanaka K, Ohinata Y, Saitou M.
Ying QL, Wray J, Nichols J, Batlle-Morera L, Doble B, Woodgett J, Cohen P, Smith A.