Reprogramming of human somatic cells to pluripotency with defined factors
Nature 451, 141-146 (10 January 2008) | doi:10.1038/nature06534; Received 16 November 2007; Accepted 10 December 2007; Published online 23 December 2007
In-Hyun Park1, Rui Zhao1, Jason A. West1, Akiko Yabuuchi1, Hongguang Huo1, Tan A. Ince2, Paul H. Lerou3, M. William Lensch1 & George Q. Daley1
Pluripotency pertains to the cells of early embryos that can generate all of the tissues in the organism. Embryonic stem cells are embryo-derived cell lines that retain pluripotency and represent invaluable tools for research into the mechanisms of tissue formation. Recently, murine fibroblasts have been reprogrammed directly to pluripotency by ectopic expression of four transcription factors (Oct4, Sox2, Klf4 and Myc) to yield induced pluripotent stem (iPS) cells. Using these same factors, we have derived iPS cells from fetal, neonatal and adult human primary cells, including dermal fibroblasts isolated from a skin biopsy of a healthy research subject. Human iPS cells resemble embryonic stem cells in morphology and gene expression and in the capacity to form teratomas in immune-deficient mice. These data demonstrate that defined factors can reprogramme human cells to pluripotency, and establish a method whereby patient-specific cells might be established in culture.
Introduction
Pluripotency is the property of cells to different into all tissues of an organism. Once differentiated from the embryonic stem cells, adult human somatic cells cannot differentiate into other types of cells. Induced pluripotency involves transfecting genes that can dedifferentiate cells into the pluripotent state. These induced pluripotent stem cells (iPSCs) have many of the same properties as Human embryonic stem cells (HSCs). Park et al. attempted to use the four main reprogramming factors to isolate iPS cells from differentiated H1-OGN cells: OCT4, SOX2, KLF4, and MYC. Undifferentiated cells of the H1-OGN line were differentiated and tested for expression of these four factors over a prolonged time period. This showed that only expression of the Myc gene persisted after 5 days of differentiation as all other expression quickly died down.
In the human ES cell line H1-OGN8, the OCT4 promoter drives expression of GFP-IRES-neo. a, Time course of differentiation of H1-OGN cells into a population of adherent fibroblasts, and subsequent expansion of a colony into a clonal fibroblast cell line (dH1cf32). The differentiated fibroblast derivatives of H1-OGN cells are morphologically indistinguishable from dermal fibroblasts cultured from an adult volunteer donor (hFib2). b, Quantitative real-time PCR demonstrates that the expression of a cohort of key pluripotency factors (OCT4, SOX2, NANOG and KLF4) is lost by the third week of differentiation, whereas expression of a fifth factor (MYC) persists.
Methods
Several experiments were performed, which involved cell culture as well as gene expression analysis. Cell culture was complicated by the use of specific factors and human ES cell medium to allow for formation of iPS colonies. Retroviral production involved introduction of OCT4, SOX2, KLF4, and c-MYC via the pMIG vector. The viral infections lasted for 24 hours and were subsequently seeded onto MEFs for five days. Chromosome counts confirmed that cell lines were diploid. Microarray analysis was performed by isolating RNA, preparation of RNA probes for microarray hybridization, and then scanned and analyzed. Assay for Teratoma formation involved injection of resuspended iPS cells into mice at a cell concentration of approximately 1 x 106 cells.
Reprogramming of human ES-cell-derived fetal fibroblasts
The differentiated cell line dH1cf was then infected with a lentiviral cocktail that had human OCT4, SOX2, Myc, and KLF4. A week later, cells were plated in HeSC culture. Two weeks after infection, small colonies were seen that were picked and expanded, which resulted in more colonies that were identical to parental H1-OGN cells. Morphology was decided to be a sufficient marker, without need for selection with G418, as seen in previous literature with murine iPS cells. Ten independent transfections of 105 dH1cf cells resulted in about 100 cells exhibiting ES-cell morphology, which is a low efficiency of about 0.1%. More interestingly, it was observed that ES-cell like colonies were still formed when Myc and Klf4 were eliminated from the viral cocktails, but the resulting efficiency was much lower.
Figure 2: Multiple cultured human primary somatic cells yield iPS cells.
a, iPS cells produced from five independent human primary cell lines form colonies with a similarly compact, ES-cell-like morphology in co-culture with mouse embryonic feeder fibroblasts (MEFs). b–f, As shown via immunohistochemistry (IHC), human iPS cell colonies express markers common to pluripotent cells, including alkaline phosphatase (AP), Tra-1-81, NANOG, OCT4, Tra-1-60, SSEA3 and SSEA4. 4,6-Diamidino-2-phenylindole (DAPI) staining indicates the total cell content per field. Fibroblasts surrounding human iPS colonies serve as internal negative controls for IHC staining. dH1f-iPS3-3 (b, from H1-OGN differentiated fibroblasts), MRC5-iPS2 (c, from MRC5 human fetal lung fibroblasts), BJ1-iPS1 (d, from neonatal foreskin fibroblasts), MSC-iPS1 (e, from mesenchymal stem cells), hFib2-iPS2 (f, dermal fibroblast from healthy adult male).
High resolution image and legend (364K)Table 1: ES-cell-like colony formation with various donor cells and reprogramming factors
Reprogramming of fetal, neonatal and adult fibroblasts
A diverse panel of human primary cells was tested with viral transfection containing the four factors (OCT4, SOX2, Myc, and KLF4) in an effort to dedifferentiate these cell lines into an iPS state. Fetal lung fibroblast cells (MRC5) and fetal skin cells (Detroit 551) formed human ES-cell like colonies after introduction of the four aforementioned factors. However, this method did not work for all of the cell lines, including neonatal foreskin fibroblasts, adult mesenchymal stem cells, and adult dermal fibroblasts.
The researchers were able to add additional factors, which eventually resulted in human ES-cell like colonies. These additional factors were suspected to be necessary for the cells to be grown in continuous cell culture and reprogramming to pluripotency. These genes were hTERT (catalytic subunit of human telomerase) and SV40 large T (anti-apoptotic). When these genes were transfected along with the four original pluripotency factors, there were some human ES-colonies, despite there still being significant cellular loss.
The isolated colonies of reprogrammed iPS cells were analyzed in several ways. Immunohistochemistry showed expression of alkaline phosphatase, Tra-1-81, Tra-1-60, SSEA3, SSEA4, OCT4, and NANOG – markers characteristic of human ES cells. Quantitative PCR studied the gene expression of the derivatives, showing that the following genes were markedly more expressed: OCT4, SOX2, NANOG, KLF4, hTERT, REX1, and GDF3. These expression levels were comparable to the parental H1-OGN human ES cells.
Although hTERT and SV40 large T genes were used in addition to the four genes to form a six-factor viral cocktail that produced iPS cells for postnatal cell lines, these genes were not expressed by the ES cell-like colonies isolated from the reprogramming. These genes may still be useful indirectly in cell culture to increase the efficiency of reprogramming, but they are not intrinsically essential for the viral transfection.
Global messenger RNA expression analysis was performed on H1-OGN cells, parental fibroblast cells, and reprogrammed iPS derivatives. Cluster plots were formed using the Perason correlation. Analysis of the scatter plots shows that there is a tighter correlation between reprogrammed cells and human ES cells than between differentiated fibroblasts and human ES cells, or differentiated fibroblasts and human iPS derivatives. Therefore, the iPS cells isolated from somatic sources are highly similar to the embryonic human stem cells at the global transcriptional level.
Finally injection of pluripotent cells into mice and subsequent teratoma formation has become a standard way to test pluripotency. The iPS cell derivatives caused cystic tumors in mice that exhibited all three primary germ layers, demonstrating the pluripotent ability of the isolated iPS cells. This result, in addition to the previous analysis involving gene expression and comparison to parental Human ES cells, suggests that the derived human ES-cell like colonies in this experiment are indeed induced pluripotent stem cells.
Figure 3: Gene expression in human iPS cells is similar to human ES cells.
a–e, Quantitative real-time PCR assay for expression of OCT4, SOX2, NANOG, MYC, KLF4, hTERT, REX1 and GDF3 in human iPS and parental cells. Individual PCR reactions were normalized against internal controls (β-actin) and plotted relative to the expression level in the parent fibroblast cell line. a, dH1f, dH1f-iPS3-3, dH1cf16-iPS-1 and dH1cf32-iPS-2 cells. b, MRC5-iPS2, MRC5-iPS12 and MRC5–iPS17. c, BJ1-iPS1. d, MSC-iPS1. e, hFib2-iPS2 and hFib2-iPS4. f, Transgene-specific PCR primers permit determination of the relative expression levels between total, endogenous (Endo) and retrovirally expressed (Transgene) genes (OCT4, SOX2, MYC and KLF4) via semi-quantitative PCR. β-Actin is shown as a positive amplification and loading control.
High resolution image and legend (257K)Figure 4: iPS cells are demethylated at the OCT4 and NANOG promoters relative to their fibroblast parent lines.
Bisulphite sequencing analysis of the OCT4 and NANOG promoters in H1-OGN human ES cells, dH1f differentiated fibroblasts, dH1f-iPS-1, dH1cf32-iPS2, as well as the MRC5 neonatal foreskin fibroblast line and its derivatives MRC5-iPS2 and MRC5-iPS19. Each horizontal row of circles represents an individual sequencing reaction for a given amplicon. White circles represent unmethylated CpG dinucleotides; black circles represent methylated CpG dinucleotides. The cell line is indicated to the left of each cluster. The values above each column indicate the CpG position analysed relative to the downstream transcriptional start site (TSS). The percentage of all CpGs methylated (% Me) for each promoter per cell line is noted to the right of each panel.
High resolution image and legend (265K)
a, A Pearson correlation was calculated and hierarchical clustering was performed with the average linkage method in H1-OGN, dH1f, dH1f-iPS3-3, dH1cf16, dH1cf-iPS cells (dH1cf16-iPS5 and dH1cf32-iPS2), MRC5, MRC5-iPS2, BJ1 and BJ1-iPS1 cells. The distance metric calculated by GeneSpring GX7.3.1 for comparisons between different cell lines is indicated above the tree lines. The fibroblast lines dH1f, dH1cf16, MRC5 and BJ1 cluster together, whereas iPS cells cluster together with the H1-OGN human ES cell line. b, Global gene expression patterns were compared between differentiated fibroblasts (dH1f, dH1cf16), reprogrammed somatic cells (dH1f-iPS3-3, MRC5-iPS2) and human ES cells (H1-OGN). Red lines indicate the linear equivalent and twofold changes in gene expression levels between the paired samples.
High resolution image and legend (131K)Figure 6: Xenografts of human iPS cells generate well-differentiated teratoma-like masses containing all three embryonic germ layers.
Immunodeficient mouse recipients were injected with human iPS cells (dH1f-iPS3-3) intramuscularly. Resulting teratomas demonstrate the following features in ectoderm, mesoderm and endoderm. Ectoderm: pigmented retinal epithelium (a), neural rosettes (b), glycogenated squamous epithelium (c); mesoderm: muscle (d), cartilage (e), bone (f); endoderm: respiratory epithelium (g). Of note, panel c contains all three germ layers: (1) glycogenated squamous epithelium, (2) immature cartilage, (3a) glandular tissue with surrounding stromal elements, and (3b) another small gland. All images were obtained from the same tumour. Tissue sections were stained with haematoxylin and eosin. Scale bar, 100 μm.
High resolution image and legend (353K)Conclusions:
This paper shows a method to induce pluripotency in several different cell types, including adult human somatic cells. The results establish feasibility in using methods similarly used for iPS reprogramming of murine cells and performing these transformations for human cells, a big step towards clinical application of these findings. OCT4 and SOX2 have been determined to be the main components of dedifferentiation, however Klf4, Myc, and other factors such as SV40 large T and hTERT can greatly increase the efficiency of this process. Before clinical success with human iPS cells can occur, there must be a developed method to avoid genetic transmutation by the lentivirus, which is nonspecific at this time. Future work would combine studies of targeted viral transfection as well as iPS factors for a predictable dedifferentiation therapy.
Critique:
This prominent study, published in Nature, is a milestone in stem cell research, because it was one of the first iPS stem cell papers that isolated the genes necessary for reprogramming to iPS state using multiple adult cell types. The four transcription factors changed to induce pluripotency were: Oct4,Sox2,Klf4, and Myc. The methods used in the study include retroviral transfection, cell culture, bisulphite genomic sequencing, and microarray analysis. The interesting phenomenon observed was that only Oct4 and Sox2 were absolutely necessary to induce pluripotency, while the other genes increased the efficiency of stem cell colony formation. This is significant, because it isolates the specific transcription factors that result in pluripotency, which is an important discovery for this type of research. Also, the study is one of the first to use human somatic cells instead of mouse somatic cells, demonstrating that iPS cell formation is possible. One problem is that the isolation of the four factors could have been more rigorous, because it was based on previous research for mouse somatic cells, so there could have been factors that were not considered. The only controls used were internal negative controls, and the researchers did not use positive controls in their microarray analysis. Finally, the quantification of iPS cell formation was not specific and is subjective. A better method of quantification should be developed instead of checking for "cell-like morphology." Overall, this paper does a good job of approaching iPS research with the significant goal of manipulating various types of adult human somatic cells.
5 comments:
This study was great in its analyses, and I am interested in how its discoveries will be used in the near future.
One thing I always wonder though is the technique of incorporating genes. The paper mentions retroviral vectors as a means of gene transfer, which may be good for efficiency, but they lack in site-specific integration, which may cause mutations. The paper also mentions homologous recombination as a technique for specifically introducing genes, but applies to meiotic cells and was not used for the 4 transcription factors. A technique for efficient site-specific transfections could further help the progress of iPS
This paper shows that it was possible to reprogram EB cells that were differentiated by just a combination of Oct4, Sox2, MYC, and KLF4. They show that only Oct4 and Sox2 were the most important reprogramming factors. They also demonstrated that these iPS cells can form teratomas, thus suggesting that the generated iPS cells are very similar to normal EB cells. Overall the analysis is great and very solid. It was interesting that they pointed out what future research would focus on- reprogramming without viral vectors- which is exactly happening among several groups now. Combined with the solid science and accurate forecast, this article is definitely worthy of being a nature paper.
This paper is very a very insightful and detailed description of the study. It is very interesting, as the area of stem cell research has been so controversial in the past, that this may prove to be an alternative way to grow and harvest stem cells. The identification of the major factors needed to create the iPS cells is very important, but the various ways of inoculating/incorporating the cells with these new genes seems like there can be a lot of room for error, as James mentioned.
This paper demonstrates the possibility of reprogramming adult cells into pluripotent stem cells, which could be similar in efficacy to embryonic stem cells. However, isn't the retrovirus non-specific in it's integration of genes? Would this affect the expression of other genes and damage the cells?
Several of the comments attribute to the merits and the deficiencies of this study. Although this is a hallmark study in iPS research, the method of retroviral transfection does have its problems. The insertion of nonspecific genes results in possibly mutated cells, which could result in cell death. As a result, we can see that there is a markedly low efficiency of 0.1%. This efficiency is even lower in the iPS cells derived from primary cells. The way to get highest yield would involve viruses that can directly target and provide specific point mutations at the desired locations to result in reprogrammed iPS cells. At this stage the use of retroviral transfection is its initial research stages and will undoubtedly be a topic of critical research over the next few years to make iPS cell technology a clinical reality.
Post a Comment