You are hereOctober 12, 2011 | Pluripotent Stem Cells
Defining the Nature of Human Pluripotent Stem Cell Progeny
From Cell Research
By Stuart P. Atkinson
Multiple studies into the differentiation capabilities of pluripotent stem cells (PSCs) such as human embryonic stem cells (hESCs) and human induced pluripotent stem cells (hiPSCs) have shown that they all have some potential to generate an array of clinically relevant cell types, even if this can vary widely across different hESC and hiPSC lines (Bock et al, Boulting et al, Feng et al and Osafune et al). Cell derivatives are often tested for specificity by a few generally well understood markers discovered through the study of natural development, but it remains to be seen if, on a global level, these derivatives are truly analogous to their natural counterparts and if their in vitro development follows what is observed in vivo. A recent study in Cell Research (Patterson et al) from the laboratory of William Lowry at the Department of Molecular, Cell and Developmental Biology, UCLA has addressed this point and overall demonstrates that hESCs and iPSCs can be differentiated towards specified derivatives that are transcriptionally similar to their developmental counterparts, but that these derivatives retain the expression of a group of genes known to play a role during very early embryonic development.
Derivatives representing all three embryonic germ layers were produced from both hESCs and iPSCs using previously established protocols for neural progenitor cells (NPCs; ectoderm) (Karumbayaram et al), hepatocytes (endoderm) (Song et al and Si-Tayeb et al), and fibroblasts (FBs; mesoderm). When PSCs were directed to generate NPCs (PSC-NPCs) it was noted that protein levels of PAX6 (lower) and NESTIN (higher) were altered in the PSC-NPCs when compared with NPCs from 16-week-old fetal brain. This was suggested to be due to the induction of a more posterior/ventral fate in the PSC-NPCs and so NPCs were isolated from 15.5-week-old fetal spinal cord and expanded under the same conditions and analysis of these cells showed similar levels of PAX6 and NESTIN to PSC-NPCs. Differences in differentiation capacity were also noted, with PSC-NPCs mostly generating neurons (Tuj1+), while the 16-week-old-tissue-derived NPCs mostly produced glia (GFAP+). This suggests that the PSC-NPCs may represent an earlier developmental time point than the NPCs derived from 16-week-old fetal tissue. When compared to counterparts made from adult liver, PSC-hepatocytes were shown to be more similar to hepatoblasts, or immature hepatocytes that populate the early developing fetal liver. PSC-hepatocytes expressed a higher level of fetal hepatic genes such as AFPand CYP3A7, and a lower level of the more mature equivalents ALBand CYP3A4, again suggesting that PSC-hepatocytes could represent an earlier developmental stage. PSC-FBs, meanwhile, showed similar RNA levels and collagen secretion profile to dermal skin cells and two fibroblast markers, CD44 and COL3A1, were also expressed in PSC-FBs at levels similar to a neonatal dermal FB line.
More detailed transcriptional analysis was then undertaken and the profile of PSC derivatives and their natural counterparts was compared. Initial analysis between undifferentiated hiPSCs and hESCs and their differentiated derivatives showed that any gene expression differences apparent in the undifferentiated states of these two cell populations were not evident upon analysis after their differentiation, while any “new” differences that arose during differentiation between that were evident between hiPSCs and hESCs progeny were not-statistically relevant. Unsupervised hierarchical clustering also showed that while PSC derivatives clustered with their respective natural counterparts, the PSC-derived cells were always distinguishable from their respective natural counterparts suggesting that these derivatives are similar but not identical to their tissue-derived counterparts. Detailed microarray analysis showed that out of 36749 probe sets, 2922 were differentially expressed between PSC-FBs and dermal/lung FBs; 4452 were differentially expressed between PSC-hepatocytes and adult hepatocytes; and 2769 were differentially expressed between PSC-NPCs and 16-week-old fetal NPCs. Of these differences, there was an overlapping number of probe sets (105, equating to 88 unique genes) demonstrating that all types of PSC derivatives share common differences with tissue-derived cells. 62 probe sets (53 genes) were upregulated in all PSC derivatives and 31 of these 62 overlap with probe sets that are highly upregulated in undifferentiated PSCs versus specified somatic cells indicating that the PSC progeny do not silence gene expression associated with pluripotency and early embryonic development. These genes include LIN28B, DPPA4, and TCF3 but not OCT4,SOX2,REX1, orNANOG. There were also 35 genes were downregulated in PSC derivatives compared to tissue-derived cells, perhaps reflecting a state of incomplete specification, regardless of the cell type generated.
The study then concentrated on the roles played by the genes that were commonly overexpressed across the PSC derivatives. LIN28B was expressed in all PSC-derivatives, whereas LIN28A was found at a high level in PSC-NPCs and PSC-hepatocytes, but not in PSC-FBs. DPPA4, LIN28A, and LIN28B were all expressed at the protein level in PSC-NPCs and PSC-Heps, but not in their tissue-derived counterparts. LIN28 acts to regulate miRNA maturation (especially let-7) and in agreement with LIN28 expression in PSC-NPCs very low relative levels of mature let-7 family members were found compared to NPCs from 16-week-old fetal brain or spinal cord which both showed high let-7 activity. The authors note that they believe theirs to be the first study to demonstrate that human PSC derivatives have high LIN28 expression and low let-7 activity. Further analyses were then undertaken in foetal tissue in order to investigate in some detail these altered gene expression patterns. DPPA4 was strongly expressed with SOX2 at 7 weeks of development along the midline of the spinal cord where the neural progenitor pool is localised, but was more weakly expressed in this region at 13 weeks. LIN28A was expressed in scattered cells in 7-week-old spinal cord, but was lost by 13 weeks, unlike LIN28B which was strongly detected in 7-week-old human spinal cord cells outside of the midline and weakly expressed in the midline progenitor cells alongside SOX2 while becoming significantly reduced, but not absent, in the spinal cord by 13 weeks of development. In the fetal liver, only LIN28B was detectable at 6.5 weeks and was undetectable by 16 weeks. Overall, LIN28 and DPPA4 expression is generally associated with earlier stages of embryonic development, and with the previous data, altogether suggests that the PSC-derivatives are similar to cells found at 7 weeks of development or earlier.
iPSCs are know to become more similar to hESCs with regards to their mRNA profile and functional characteristics upon extended growth in culture, possibly due to positive feedback mechanisms altering epigenetic profiles. This data prompted the researchers to investigate the role of extended culture in PSC-derivatives and whether this could allow the cells to become more like their tissue-derived counterparts and additionally if even younger foetal samples would be more similar at the mRNA level to the hESCs and hiPSCs. Upon extended culture of PSC-NPCs, both a reduction in LIN28A and LIN28B mRNA and protein and upregulation of let-7 family members was observed, while DPPA4 was unchanged. This was associated with a small increase in transcriptomic similarity between PSC-NPCs and fetal-derived NPCs. Further, a small but statistically significant number of the original 2723 probe sets were “corrected” upon extended passaging, including LIN28A and LIN28B, paralleling the changes observed in iPSCs during prolonged culture. Extended culture also led to an increase in PSC-NPCs gliogenic capability (from ~1% to ~15%) but not to a level consistent with their tissue-derived counterparts. Indeed, together the data collected did not support the hypothesis that extended culture of NPCs can generate cells equivalent to their tissue derived counterparts. Comparisons of PSC-NPCs with additional fetal samples from 6.5 to 8 weeks of development led to a dramatic increase in transcriptomic similarity. A study which examined gene expression across whole human embryos from 3-5 weeks of development (Yang et al) was used to compare probe sets that were differentially expressed between 3 and 5 week embryos and those found to be different between PSCs and their tissue-derived counterparts. Significant overlap was found between these two sets of genes; with LIN28A being the most differentially expressed probe set between 3 and 5 week embryos. Detailed analysis showed that a significant number of genes that are normally downregulated between 3-5 weeks of development also appear to distinguish PSC derivatives from their tissue-derived counterparts, further suggesting that PSC derivatives might accurately recapitulate cells found prior to 6 weeks of development.
These findings raise some interesting points. Do current differentiation strategies properly recapitulate in vivo development? Should we be studying in vivo development further and applying these strategies in vitro for study instead of moving ahead with ambitious clinical trials involving possibly functionally aberrant stem cell derivatives? The limitations of in vitro differentiation are well documented and accepted, and so studies like this further confirm this notion. If we do go ahead with current PSC-derived cells in clinical trials, do we know what affect the aberrantly expressed genes associated with early embryo development will have on long-term functionality? For example, the expression of LIN28 has been previously linked to cancer (Viswanathan et al) and one could argue that the expression of early development genes in adult tissues is a possible threat to the patient. Perhaps further studies into epigenetic differences may also find yet more differences between the in vivo and in vitro cell derivatives.
Bock C, Kiskinis E, Verstappen G, et al.
Reference Maps of human ES and iPS cell variation enable high-throughput characterization of pluripotent cell lines.
Cell. 2011 Feb 4;144(3):439-52.
Boulting GL, Kiskinis E, Croft GF, et al.
A functionally characterized test set of human induced pluripotent stem cells.
Nat Biotechnol. 2011 Mar;29(3):279-86.
Fang H, Yang Y, Li C, et al.
Transcriptome analysis of early organogenesis in human embryos.
Dev Cell. 2010 Jul 20;19(1):174-84.
Feng Q, Lu SJ, Klimanskaya I, et al.
Hemangioblastic derivatives from human induced pluripotent stem cells exhibit limited expansion and early senescence.
Stem Cells. 2010 Apr;28(4):704-12.
Karumbayaram S, Novitch BG, Patterson M, et al.
Directed differentiation of human-induced pluripotent stem cells generates active motor neurons.
Stem Cells. 2009 Apr;27(4):806-11.
Osafune K, Caron L, Borowiak M, et al.
Marked differences in differentiation propensity among human embryonic stem cell lines.
Nat Biotechnol. 2008 Mar;26(3):313-5.
Patterson M, Chan DN, Ha I, et al.
Defining the nature of human pluripotent stem cell progeny.
Cell Res. 2011 Aug 16. doi: 10.1038/cr.2011.133.
Song Z, Cai J, Liu Y, et al.
Efficient generation of hepatocyte-like cells from human induced pluripotent stem cells.
Cell Res. 2009 Nov;19(11):1233-42.
Si-Tayeb K, Noto FK, Nagaoka M, et al.
Highly efficient generation of human hepatocyte-like cells from induced pluripotent stem cells.
Hepatology. 2010 Jan;51(1):297-305.
Viswanathan SR, Powers JT, Einhorn W, et al.
Lin28 promotes transformation and is associated with advanced human malignancies.
Nat Genet. 2009 Jul;41(7):843-8.