Reply to |[ldquo]|On the cloning of animals from terminally differentiated cells|[rdquo]|
Xiangzhong Yang,Tao Cheng,Li-Ying Sung,Shaorong Gao,Hongmei Shen,Hui Yu,Yifang Song,Sadie L Smith,David P Tuck,Kimiko Inoue,Sherman M Weissman
DOI: https://doi.org/10.1038/ng0207-137
IF: 30.8
2007-01-01
Nature Genetics
Abstract:Hochedlinger and Jaenisch accurately summarized our paper's claims1 and then made three criticisms: (i) that our conclusions are not novel in light of previous work; (ii) that because our differentiated cell population is 99.4% (not 100%) pure, the clones we obtained from this population could be derived from a rare subset of undifferentiated cells and (iii) that our data are insufficient to compare the efficiency of cloning from differentiated versus undifferentiated adult stem cells. We address these points in turn. Our study addresses the efficiency of reprogramming somatic nuclei at different differentiation stages. We are surprised that Hochedlinger and Jaenisch question the novelty of our findings, given that they themselves, in a review published a few months ago2, identified this as an important question: "Whether the genomes of adult stem cells are similar to ES cells in that they are easier to reprogram than the genomes of terminally differentiated cells is an open question." Several excellent studies3, 4, 5, all of which we cited, have reported the derivation of mice by nuclear transfer from terminally differentiated cells. However, they differ from ours in that none of those studies addressed the efficiency of reprogramming of somatic nuclei at different stages of differentiation and none succeeded in making a cloned mouse directly. They had to resort to a two-step process where a donor nucleus from a differentiated cell was used to make a cloned blastocyst from which ES cells were derived and used to make a fetal-placenta chimera via tetraploid complementation procedures3, 4, 5 or for a second round of nuclear transfer4. These studies (with peer commentaries6, 7, 8, 9) differ from ours, as shown in Figure 1. Our studies are therefore the first unambiguous demonstration of nuclear totipotency of terminally differentiated cells by direct nuclear transfer. We found unambiguous, convincing morphological evidence that our population of purified differentiated granulocytes was 99.4% pure after FACS for Gr-1high expression, a different method1 from the one cited by Hochedlinger and Jaenisch. They suggest that the two cloned mice obtained from this population could have come from the 0.6% of cells that are undifferentiated, but we consider this extremely unlikely. Our estimate of 99.4% purity was based on morphological examination of 1,000 cells by two independent observers. The remaining 0.6% of observed cells were myelocytes or metamyelocytes, which are also lineage-committed terminally differentiated cells; we did not observe any hematopoietic stem cells or monocytes (Gr-1low), although monocytes are fully differentiated cells, too. Hochedlinger and Jaenisch argue that morphologically based identification is inherently unreliable. This is not true, as it is the gold standard for recognizing granulocytes in the field of hematology. The chance that the clones were contaminated by primitive HSCs, which constitute only 1 10-5 to 5 10-5 bone marrow nucleated cells even before sorting, is almost zero. We reported that cloning efficiency increases over the differentiation hierarchy in the hematopoietic cell lineage. Hochedlinger and Jaenisch did not agree with our conclusion because of their trivial explanation for our result: namely, that the low efficiency of HSC nuclei might be because (i) the nuclei of HSCs are more easily damaged by nuclear transfer, FACS or the freeze-thaw procedures or (ii) blastocyst development is not an acceptable measure for judging cloning efficiency, and generation of viable animals or ES cell lines is required for comparing cloning efficiency. This is an ad hoc criticism that could be leveled against any result, but we consider it unlikely as an explanation of our findings. We compared cloning efficiency of four types of hematopoietic cells with varying degrees of differentiation along a single lineage: long-term repopulating HSCs, short-term repopulating HSCs, hematopoietic progenitors and differentiated granulocytes, all from the same animal tissue and all isolated, purified, and frozen and thawed using the exact same standard method1. They argue the possibility of some mechanical damage to HSCs during FACS or freezing and thawing, which is very unlikely. Our validations showed that the FACS-isolated HSCs were highly functional in engraftment or culture assays1 and that HSCs either freshly isolated or used after freezing and thawing had similarly poor cloning efficiency in both our studies1, 10, suggesting that the freeze-thaw procedure had no impact. The efficiency of cloning, defined as the percentage of reconstructed embryos that reach morula/blastula stage, shows a consistent trend: efficiency becomes progressively higher as differentiation proceeds, with more than an eightfold difference in efficiency between the least and the most differentiated cells (4% versus 35%). This conclusion is based on over 3,000 nuclear transfer injections. Hochedlinger and Jaenisch argue that the efficiency of development to the blastocyst stage is not an acceptable measure of cloning efficiency. We disagree with their opinion, as this is a commonly used measure in most animal cloning reports, and we note that they3, 4 as well as others that they cited5, 10 have used similar measures in previous studies. While we agree that live births would be the most definitive measurement for totipotency, this is not feasible given the numbers that would be needed. We did transfer some cloned blastocysts derived from HSCs (n = 28, stated as "data not shown"), but none went to term. In our two studies, we failed to generate clones with B6D2F1 HSCs1, 10, although we (K.I.) produced 2 cloned pups from 302 (B6129)F1 HSC–derived embryos10. As many know, the 129 genome helps efficient genomic reprogramming, and it is not reasonable to do a direct comparison between different mouse strains, as shown in their table. We have generated cloned ES cell lines from HSCs (data not shown). The focus of this study was not to generate ES cells or 'clonal' mice via two-step cloning as previously reported3, 4, 5, but to directly compare the developmental potential of differentiated cells versus adult stem cells from the same tissue lineage using the conventional nuclear transfer method that led to the creation from adult somatic cells of cloned animals like Dolly. Our surprising finding provides unequivocal evidence that efficiency of conventional methods of cloning increases with the differentiation hierarchy within a hematopoietic cell lineage, although more studies are required to determine whether this trend is reversed within other cell lineages or under modified conditions.