Answers are grounded in this paper and cite their sources.
Dear Editor, We previously demonstrated (Jiang et al., 2002) that rodent multipotent adult progenitor cells (MAPC) can self-renew long-term while maintaining multilineage differentiation capacity. Rodent MAPC express a number of pluripotency-related transcription factors (TF) including Oct4 and Rex1 but not Nanog and Sox2, two other TF known to play a significant role in the maintenance of the pluripotency of embryonic stem cells (ESC) (Ulloa-Montoya et al., 2007). However, rodent MAPC express several TF, including Gata4, Gata6, Sox7 and Sox17, typically expressed in the nascent hypoblast of the developing inner cell mass (ICM) (Nichols and Smith, 2011) and in the recently described rat extrambryonic endodermal precursor cells (rXEN-P), which are isolated from blastocyst (Debeb et al., 2009). We derived in 4/12 independent isolations one or more rMAPC lines, by culturing rat BM cells in rMAPC medium (rMAPC isolation scheme, Supplementary Figure S1). After 4 weeks of culture, BM cells were depleted of CD45+ cells and 2–8 weeks later, clusters of refractile and small cells appeared, which became the preponderant cell type within 10 days (Figure 1A). Nearly all cells from the established lines expressed Oct4, Gata4, Gata6, Sox7 and Sox17 transcripts and proteins (Figure 1B and Supplementary Figure S2A and B), as well the surface markers SSEA1 and CD31 (Figure 1C and Supplementary Figure S2C), both markers of the early ICM. Similarities between rMAPC, in vitro generated from BM, and rHypoSC, derived from blastocysts. (A) Morphology of BM cells during the initial high-density culture (left picture); one month later, after CD45 column depletion (middle picture) and of the 3C3 rMAPC clone (right picture) (for isolation scheme, see Supplementary Figure S1). (B) Immunostainings on 3C3 rMAPC for Oct4, Gata4, Gata6, Sox7, Sox17. Scale bars represent 50 μm. (C) FACS analysis for CD31 and SSEA1 on cells obtained after CD45 column depletion (left plot), on the 3C3 rMAPC clone (middle plot) and on the Fi1 rHypoSC clone. (D) Immunohistology for Oct4 on cells obtained after CD45 column depletion, and on the 3C3 rMAPC line. (E) RT-qPCR analysis for the nascent hypoblast gene signature before, during and after the appearance of the refractile and small rMAPC. ΔCT for the indicated genes is shown, after normalization with the housekeeping gene Gapdh, at different timepoints after CD45 depletion (dx). (F) Sequential outgrowth of rHypoSC from a rat blastocyst. (G) Immunostainings on Fi1 rHypoSC line for Oct4, Gata4, Gata6, Sox7, Sox17. Scale bars represent 50 μm. (H) RT-qPCR comparison for the hypoblast gene expression levels between rMAPC (CL19, rMAPC-1 and 3C3 line) and rHypoSC (Fi1, Fi2, Wk7 and Wk8 line). Shown are mean ± SEM of the ΔCT for the indicated genes, normalized using Gapdh as housekeeping gene. (I) Microarray comparison. Heatmap of the Pearson correlation is shown. The coefficient of determination (R2) between the average expression profiles of mesenchymal stem cells (MSC), rMAPC and rHypoSC cell lines is also shown on the right. (J) Chimeras (E7.0) generated after morula aggregation with rMAPC (CL19) and rHypoSC (Fi1). Scale bars represent 100 μm. Although rMAPC lines express Oct4, previous studies (Lengner et al., 2007) demonstrated that Oct4+ cells cannot be detected in adult mouse tissues and that Oct4 is not required for postnatal tissue homeostasis. Based on Lengner's findings, we hypothesized that the rMAPC phenotype could be the result of a culture-induced reprogramming. We therefore analysed BM-cultures during 2 independent rMAPC isolations before, during and after the appearance of the refractile and small cells. We could not identify any Oct4+ or SSEA1+/CD31+ cells in more than one million cells analyzed after CD45+ cells depletion (Figure 1C and D, and Supplementary Figure S2D and E), several weeks before the appearance of the refractile cells positive for these markers. RT-qPCR analysis further demonstrated that acquisition of the typical rMAPC morphology was associated with >1000-fold increase in expression of Oct4 and the typical hypoblast gene transcripts (Figure 1E and Supplementary Figure S2F). Although some rMAPC lines had karyotypical abnormalities (Supplementary Figure S2G and Supplementary Data), some lines did not, suggesting that the rMAPC phenotype is not induced by a specific translocation, duplication and/or deletion. These studies demonstrate that rMAPC do not exist in BM and that this hypoblast phenotype is acquired upon prolonged in vitro culture. rMAPC may represent a rare event of in vitro reprogramming, resembling what has been observed during spermatogonial stem cell (Guan et al., 2006; Kanatsu-Shinohara et al., 2008; Ko et al., 2009) and epiblast stem cell (Bao et al., 2009) de-differentiation to ESC-like cells, when cultured under ESC conditions. Because the gene expression pattern of rMAPC (Ulloa-Montoya et al., 2007) and rXEN-P cells (Debeb et al., 2009) is highly similar, we next asked whether BM cells were reprogrammed to a hypoblast/extraembryonic progenitor fate. To investigate this, we tested whether rMAPC could be cultured under rXEN-P conditions and vice versa. When established rXEN-P lines were cultured under rMAPC conditions for 1–2 passages, they grew dispersed, acquiring the typical rMAPC morphology (Supplementary Figure S3A). rXEN-P cells became homogeneously Oct4+/Gata4+ and the percentage of SSEA1+ cells increased (Supplementary Figure S3B and E). RT-qPCR revealed that no differences in RNA expression for hypoblast genes could be detected in XEN-P lines, once cultured in MAPC conditions, except for higher levels of Sox17 and lower levels of Tmprss2 (Supplementary Figure S3D). By contrast, when rMAPC were cultured in XEN-P medium on rat embryonic feeders, typical XEN-P colonies were generated, i.e. Oct4−/Gata4+ epithelioid cells with a rim of loosely attached small refractile cells that are Oct4+/Gata4+ (Supplementary Figure S3F and G). Moreover, the percentage of SSEA1+ cells decreased (Supplementary Figure S3J); consistently, RT-qPCR revealed a decrease in Sox17 and an increase in Tmprss2 (Supplementary Figure S3I). Cell doubling time of rMAPC or rXEN-P cells cultured in MAPC conditions was slightly faster than in XEN-P conditions (Supplementary Figure S3A and F). Therefore, rMAPC culture conditions supported the feeder-free growth of established XEN-P clones in a more homogenous and immature state. To further define the relationship between rMAPC, XEN-P and typical XEN cells, rMAPC and rXEN-P cells were also cultured under standard XEN conditions (Kunath et al., 2005) without exogenous LIF (Supplementary Figure S4A). rXEN-P and rMAPC cells in XEN conditions, formed extraembryonic endodermal colonies with significantly lower proliferation rate (Supplementary Figure S4B and E). Expression of hypoblast gene transcripts did not change significantly when rXEN-P were cultured under XEN conditions, except for a decrease in Oct4 levels (Supplementary Figure S4D), while differences were more prominent for rMAPC (decrease in Sox17 and increase of Hnf4a and Tmprss2; Supplementary Figure S4G). SSEA3, a marker of primitive/visceral endoderm, could be detected in a subpopulation of cells grown under MAPC or XEN-P conditions (Supplementary Figure S3C and H), but not under XEN conditions (Supplementary Figure S4C and F). Expression of laminin B and collagen type IV (two proteins of the basement membrane, produced by extraembryonic endoderm) was detected irrespective of the culture conditions (Supplementary Figures S3C and H, and S4C and F). However, under the MAPC culture conditions, the laminin B and collagen 4 staining was more punctuated and less abundant, and it was mostly still cell-associated, as may be expected from a presecretory, less differentiated state. We next tested if rMAPC conditions would allow the isolation of homogeous populations of cells with hypoblast/extraembryonic endoderm features directly from rat blastocysts. When blastocysts were plated individually under rMAPC conditions, small and refractile cells emerged in 2–10 days (Figure 1F). These cells were morphologically identical to BM-derived rMAPC and were easily expanded (5–10 million cells in two weeks). More than 70% of isolation attempts were successful. Cell surface phenotype, immunophenotype, expression of hypoblast gene signature (Figure 1C, G and H), and cell doubling time of the blastocyst derived cell lines were highly similar to the BM-derived rMAPC (Supplementary Figure S5G). Immunostaining of the initial cell outgrowth demonstrated that new lines stain positive for Oct4 and Gata4, but were negative for Nanog (Supplementary Figure S5A and B). Differently from rXEN-P, which emerge much slower on feeders through an Oct4− initial intermediate (Debeb et al., 2009), new isolates expressed Oct4 even immediately after isolation. Because of their resemblance to the E3.5 nascent hypoblast, we termed the new lines rat hypoblast stem cells (rHypoSC). Microsatellite and Y chromosome PCR analysis ascertained that new rHypoSC lines were not the result of cross contamination with previously isolated rMAPC or rXEN-P lines (Supplementary Figure S5C and D). As was seen for rMAPC, some, but not all, rHypoSC lines displayed cytogentical abnormalities (Supplementary Figure S5E and Table S1). The new lines were positive for SSEA3, laminin B and collagen type IV (Supplementary Figure S5F). rHypoSC formed the typical extraembryonic endodermal colonies, when switched to XEN-P or XEN medium (Supplementary Figure S5G). Culture under XEN conditions also resulted in a longer cell doubling time, decreased expression of Oct4 and Sox17 and increased expression of Hnf4a and Tmprss2 (Supplementary Figure S5G and H), as also shown for rMAPC and rXEN-P. We then compared the growth factor requirements of rMAPC and rHypoSC. Omission of LIF prevented colony formation completely with loss of Oct4 and Tbx3 expression and acquisition of Afp, Hnf4a, and Tmprss2 expression, consistent with spontaneous differentiation towards visceral (VE) and parietal endoderm (PE). Omission of PDGF resulted in a dramatic reduction in colony formation (Supplementary Figure S6A and B). Pyrosequencing of 4 CpG islands in the proximal promoter of Oct4 demonstrated minimal methylation in the rMAPC and rHypoSC lines (Supplementary Figure S6C), while the promoter was methylated in rat tail fibroblasts, as well as in CD45 column-depleted cells, from the same isolation wherein subsequently rMAPC appeared. We compared the global gene expression profile of three rMAPC, three rHypoSC lines with a mesenchymal stem cell (MSC)-like cell line, isolated and cultured in rMAPC medium but not expressing Oct4 and the hypoblast signature (Ulloa-Montoya et al., 2007). Principal component analysis (PCA) and Pearson correlation heatmap (Figure 1I) demonstrated the similarities between the rMAPC and rHypoSC. Despite the differences in origin and time required for their isolation, rMAPC and rHypoSC lines are located in the same position in the first component (X-axis) by PCA. However, differences can be detected in the second component (Y-axis) (Supplementary Figure S6D, and Tables S2 and S3). To address the functional properties of rMAPC and rHypoSC, multiple cell lines were differentiated to smooth muscle and hepatocyte. Similar levels of marker gene expression were seen for all rMAPC and rHypoSC lines (Supplementary Figure S7). As rXEN-P cells were shown to generate yolk-sac like tumors and contribute to the PE and VE after morula aggregation, we tested if rMAPC and rHypoSC lines had similar in vivo developmental potential. When 2 rMAPC and 2 rHypoSC lines were transplanted under the skin of immunodeficient mice, yolk-sac like tumors, staining positive for AFP, periodic acid shiff, laminin and also containing Pan-cytokeratin positive epithelial structures, were detected two months after transplantation (Supplementary Figure S8). Finally, we also aggregated murine morulas with GFP-labeled rMAPC and rHypoSC (Figure 1J). Although the degree of chimerism was modest, we consistently detected descendants in the parietal and visceral yolk sac, but never in the embryo proper (Supplementary Figure S9A–E). Contribution to VE was more sporadic, and was observed in the most-distal VE (Supplementary Figure S9J–M). Occasionally, contribution to the trophoblast lineage was observed with Cl19 rMAPC line, although GFP+ cells were not yet transformed to giant cells (Supplementary Figure S9F–I). We previously demonstrated that one line of mouse (m)MAPC contributed to the embryo proper (Jiang et al., 2002). Recent studies demonstrate that cells of nascent hypoblast are even more plastic than their epiblast-counterpart and can contribute to all three lineages of the blastocyst until peri-implantation (Grabarek et al., 2012). This plasticity may explain the contribution of one rMAPC line to the trophoblast in this study, while another mMAPC line contributed to the embryo. Similarly, this may explain why cells with a hypoblast fate, derived from blastocysts or induced in BM cells, differentiate to hepatoyctes. We recently published that hepatocyte-like differentiation from rMAPC occurs likely via a primitive streak stage (Roelandt et al., 2010). This in vitro conversion mirrors the well-known behavior of ESC, which rarely contribute to the extraembryonic endoderm in vivo (Beddington and Robertson, 1989) but easily cross lineage borders and differentiate into extraembryonic endoderm in vitro (Doetschman et al., 1985). Moreover, the study of Kwon and colleagues has shown that AfpGFP positive VE cells contribute, at least transiently, to E8.5 hindgut and foregut (Kwon et al., 2008); therefore, during embryo development, such strict segregation between embryonic and extraembryonic fates may also not be true. In conclusion, by applying the modern repertoire used to define cellular identity, we demonstrate that rMAPC are highly similar to rXEN-P and to the newly isolated rHypoSC lines. In contrast to rHypoSC, that can be isolated in one week, rMAPC emerge abruptly and at low frequency, after several weeks of culture of BM in rMAPC medium, and appear to be a conversion of a rBM cell to a cell with nascent hypoblast properties. The mechanism underlying the conversion remains unknown. The ability to quickly and efficiently isolate homogenous rHypoSC populations from rat blastocysts, which contribute to the PE and VE, suggests the existence of a stable hypoblast entity during the ICM segregation. HypoSC will facilitate the further characterisation of the extraembryonic endodermal compartment, which seems to be more ‘plastic’ than previously believed and may also contribute to the definitive endoderm of the embryo proper. Therefore, HypoSC may not only be a tool for developmental biology studies but also a source to obtain cell types for regenerative medicine. [Supplementary material is available at Journal of Molecular Cell Biology online. We acknowledge the VIB microarray facility of the KULeuven for expert help and advise with the microarray experiments. We thank the people of the Stem Cell Institute Leuven for critical reading and for the continuous support with cell culture. C.M.V. was supported by an Odysseus grant from FWO, Flanders, by a CoE K.U.Leuven award, FP6-STROKEMAP from the EC, IWT-SBO-HEPSTEM Flanders. C.M.V. and A.Z. are supported by GOA/11/012 from K.U.Leuven, B.B. was supported by KOSEF (0000179) and a KUL visiting professor fellowship. A.L.N., V.R. and P.R. were supported by an IWT grant, Flanders. M.G. and A.C. were supported by FWO, Flanders. The cytogenetic analyses were performed in the Cytogenetics Core Laboratory at the University of Minnesota with support from the comprehensive Masonic Cancer Center NIH Grant #P30 CA077598-09.]
No takes yet. Share an insight, caveat, or question.
Nigro et al. (2012) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: