- Articles On Information

ARTICLE

Stimulus-triggered fate conversion of somatic cells into pluripotency

Haruko Obokata1,2,3, Teruhiko Wakayama3{, Yoshiki Sasai4, Koji Kojima1, Martin P. Vacanti1,5, Hitoshi Niwa6, Masayuki Yamato7 & Charles A. Vacanti1

doi:10.1038/nature12968

Here we report a unique cellular reprogramming phenomenon, called stimulus-triggered acquisition of pluripotency (STAP), which requires neither nuclear transfer nor the introduction of transcription factors. In STAP, strong external stimuli such as a transient low-pH stressor reprogrammed mammalian somatic cells, resulting in the generation of plu- ripotent cells. Through real-time imaging of STAP cells derived from purified lymphocytes, as well as gene rearrange- ment analysis, we found that committed somatic cells give rise to STAP cells by reprogramming rather than selection. STAP cells showed a substantial decrease in DNA methylation in the regulatory regions of pluripotency marker genes. Blastocyst injection showed that STAP cells efficiently contribute to chimaeric embryos and to offspring via germline transmission. We also demonstrate the derivation of robustly expandable pluripotent cell lines from STAP cells. Thus, our findings indicate that epigenetic fate determination of mammalian cells can be markedly converted in a context-dependent manner by strong environmental cues.

In the canalization view of Waddington’s epigenetic landscape, fates of somatic cells are progressively determined as cellular differentiation proceeds, like going downhill. It is generally believed that reversal of differentiated status requires artificial physical or genetic manipulation of nuclear function such as nuclear transfer1,2 or the introduction of multiple transcription factors3. Here we investigated the question of whether somatic cells can undergo nuclear reprogramming simply in response to external triggers without direct nuclear manipulation. This type of situation is known to occur in plants—drastic environmental changes can convert mature somatic cells (for example, dissociated carrot cells) into immature blastema cells, from which a whole plant structure, including stalks and roots, develops in the presence of auxins4. A chal- lenging question is whether animal somatic cells have a similar potential that emerges under special conditions. Over the past decade, the pres- ence of pluripotent cells (or closely relevant cell types) in adult tissues has been a matter of debate, for which conflicting conclusions have been reported by various groups5–11. However, no study so far has proven that such pluripotent cells can arise from differentiated somatic cells.

Haematopoietic cells positive for CD45 (leukocyte common antigen) are typical lineage-committed somatic cells that never express pluripotency- related markers such as Oct4 unless they are reprogrammed12,13. We therefore addressed the question of whether splenic CD451 cells could acquire pluripotency by drastic changes in their external environment such as those caused by simple chemical perturbations.

Low pH triggers fate conversion in somatic cells

CD451 cells were sorted by fluorescence-activated cell sorting (FACS) from the lymphocyte fraction of postnatal spleens (1-week old) of C57BL/6 mice carrying an Oct4-gfp transgene14, and were exposed to various types of strong, transient, physical and chemical stimuli (described below). We examined these cells for activation of the Oct4 promoter after culture for several days in suspension using DMEM/F12 medium supplemented with leukaemia inhibitory factor (LIF) and B27

(hereafter called LIF1B27 medium). Among the various perturbations, we were particularly interested in low-pH perturbations for two reasons. First, as shown below, low-pH treatment turned out to be most effective for the induction of Oct4. Second, classical experimental embryology has shown that a transient low-pH treatment under ‘sublethal’ conditions can alter the differentiation status of tissues. Spontaneous neural conver- sion from salamander animal caps by soaking the tissues in citrate-based acidic medium below pH 6.0 has been demonstrated previously15–17.

Without exposure to the stimuli, none of the cells sorted with CD45 expressed Oct4-GFP regardless of the culture period in LIF1B27 medium. In contrast, a 30-min treatment with low-pH medium (25-min incuba- tion followed by 5-min centrifugation; Fig. 1a; the most effective range was pH 5.4–5.8; Extended Data Fig. 1a) caused the emergence of sub- stantial numbers of spherical clusters that expressed Oct4-GFP in day-7 culture (Fig. 1b). Substantial numbers of GFP1 cells appeared in all cases performed with neonatal splenic cells (n 5 30 experiments). The emer- gence of Oct4-GFP1 cells at the expense of CD451 cells was also observed by flow cytometry (Fig. 1c, top, and Extended Data Fig. 1b, c). We next fractionated CD451 cells into populations positive and negative for CD90 (T cells), CD19 (B cells) and CD34 (haematopoietic progenitors18), and subjected them to low-pH treatment. Cells of these fractions, including T and B cells, generated Oct4-GFP1 cells at an efficacy com- parable to unfractionated CD451 cells (25–50% of surviving cells on day 7), except for CD341 haematopoietic progenitors19, which rarely produced Oct4-GFP1 cells (,2%; Extended Data Fig. 1d).

Among maintenance media for pluripotent cells20, the appearance of Oct4-GFP1 cells was most efficient in LIF1B27 medium, and did not occur in mouse epiblast-derived stem-cell (EpiSC) medium21,22 (Extended Data Fig. 1e). The presence or absence of LIF during days 0–2 did not substantially affect the frequency of Oct4-GFP1 cell gen- eration on day 7 (Extended Data Fig. 1f), whereas the addition of LIF during days 4–7 was not sufficient, indicating that LIF dependency started during days 2–4.

1Laboratory for Tissue Engineering and Regenerative Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115, USA. 2Laboratory for Cellular Reprogramming, RIKEN Center for Developmental biology, Kobe 650-0047, Japan. 3Laboratory for Genomic Reprogramming, RIKEN Center for Developmental biology, Kobe 650-0047, Japan. 4Laboratory for Organogenesis and Neurogenesis, RIKEN Center for Developmental biology, Kobe 650-0047, Japan. 5Department of Pathology, Irwin Army Community Hospital, Fort Riley, Kansas 66442, USA. 6Laboratory for Pluripotent Stem Cell Studies, RIKEN Center for Developmental biology, Kobe 650-0047, Japan. 7Institute of Advanced Biomedical Engineering and Science, Tokyo Women’s Medical University, Tokyo 162-8666, Japan. {Present address: Faculty of Life and Environmental Sciences, University of Yamanashi, Yamanashi 400-8510, Japan.

30 JANUARY 2014 | VOL 505 | NATURE | 641

d0 d1 d2 d3

Lymphocyte fraction

FACS sorting

Centrifugation

Plating

d0 d7

Spleen (Oct4-gfp)

CD45+ 37 °C, 25 min

Remove supernatant

DMEM/F12 medium (B27+LIF)

pH 5.7

Resuspend in medium

High magnification

Control

Low-pH-treated cells

Oct4-GFP+ Oct4-GFP–

Low-pH-treated cells d7

Non-treated cells d7

CD45

STAP

CD45

Oct4-GFP+ Forward scattering

4 6 8 10 Distance(μm)

CD45+

ES i GL

600 500 400 300 200 100

d1 d2 d3

300 200

d4 d5 d6 100 0

d0 d1 d2 d3 d4 d5 d6 d7

1 2 3

4 5

Figure 1 | Stimulus-triggered conversion of lymphocytes into Oct4-GFP1 cells. a, Schematic of low-pH treatment. b, Oct4-GFP1 cell clusters appeared in culture of low-pH-treated CD451 cells (middle; high magnification, right) on day 7 (d7) but not in culture of control CD451 cells (left). Top: bright-field view; bottom, GFP signals. Scale bar, 100 mm. c, FACS analysis. The x axis shows CD45 epifluorescence level; y axis shows Oct4-GFP level. Non-treated, cultured in the same medium but not treated with low pH. d, GFP1 (green) and GFP2 (yellow) cell populations (average cell numbers per visual field; 310 objective lens). n 5 25; error bars show average 6 s.d. e, Snapshots of live imaging of culture of low-pH-treated CD451 cells (Oct4-gfp). Arrows indicate cells that started expressing Oct4-GFP. Scale bar, 50 mm. f, Cell size reduction in

Most of the surviving cells on day 1 were still CD451 and Oct4-GFP2. On day 3, the total cell numbers were reduced to between one-third to one-half of the day 0 population (Fig. 1d; see Extended Data Fig. 1g, h for apoptosis analysis), and a substantial number of total surviving cells became Oct4-GFP1 (Fig. 1d), albeit with relatively weak signal inten- sity. On day 7, a significant number of Oct4-GFP1CD452 cells (one-half to two-thirds of total surviving cells) constituted a distinct population from the Oct4-GFP2CD452 cells (Fig. 1c, top, day 7, and Fig. 1d). No obvious generation of Oct4-GFP1CD452 populations was seen in non- treated CD451 cells cultured similarly but without low-pH treatment (Fig. 1c, bottom).

Low-pH-treated CD451 cells, but not untreated cells, gradually turned on GFP signals over the first few days (Fig. 1e, Supplementary Videos 1 and 2 and Extended Data Fig. 2a), whereas CD45 immunoreactivity became gradually reduced in the cells that demonstrated Oct4-GFP expression (Fig. 1f and Extended Data Fig. 2b). By day 5, the Oct4-GFP1 cells attached together and formed clusters by accretion. These GFP1 clusters (but not GFP2 cells) were quite mobile and often showed cell processes on moving (Supplementary Video 1).

The Oct4-GFP1 cells demonstrated a characteristic small cell size with little cytoplasm and also showed a distinct fine structure of the nucleus compared with that of parental CD451 lymphocytes (Fig. 1g). The Oct4-GFP1 cells on day 7 were smaller than non-treated CD451 cells (Fig. 1g, h and Extended Data Fig. 2c) and embryonic stem (ES) cells (Fig. 1h), both of which are generally considered to be small in

642 | NATURE | VOL 505 | 30 JANUARY 2014

low-pH-treated CD451 cells on day 1 before turning on Oct4-GFP without cell division on day 2. In this live imaging, cells were plated at a half density for easier viewing of individual cells. Scale bar, 10 mm. g, Electron microscope analysis. Scale bar, 1 mm. h, Forward scattering analysis of Oct4-GFP2CD451 cells (red) and Oct4-GFP1CD452 cells (green) on day 7. Blue line, ES cells.

i, Genomic PCR analysis of (D)J recombination at the Tcrb gene. GL is the size of the non-rearranged germline type, whereas the smaller ladders correspond to the alternative rearrangements of J exons. Negative controls, lanes 1, 2; positive controls, lane 3; FACS-sorted Oct4-GFP1 cells (two independent preparations on day 7), lanes 4, 5.

size. The diameter of low-pH-treated CD451 cells became reduced during the first 2 days, even before they started Oct4-GFP expression (Fig. 1f), whereas the onset of GFP expression was not accompanied by cell divisions. Consistent with this, no substantial 5-ethynyl-29- deoxyuridine (EdU) uptake was observed in the Oct4-GFP1 cells after the stressor (Extended Data Fig. 2d).

The lack of substantial proliferation argues against the possibility that CD452 cells, contaminating as a very minor population in the FACS-sorted CD451 cells, quickly grew and formed a substantial Oct4- GFP1 population over the first few days after the low-pH treatment. In addition, genomic rearrangements of Tcrb (T-cell receptor gene) were observed in Oct4-GFP1 cells derived from FACS-purified CD451 cells and CD901CD451 T cells (Fig. 1i, lanes 4, 5, and Extended Data Fig. 2e–g), indicating at least some contribution from lineage-committed T cells. Thus, Oct4-GFP1 cells were generated de novo from low-pH- treated CD451 haematopoietic cells by reprogramming, rather than by simple selection of stress-enduring cells23.

Low-pH-induced Oct41 cells have pluripotency On day 7, the Oct4-GFP1 spheres expressed pluripotency-related marker proteins22 (Oct4, SSEA1, Nanog and E-cadherin; Fig. 2a) and marker genes (Oct4, Nanog, Sox2, Ecat1 (also called Khdc3), Esg1 (Dppa5a), Dax1 (Nrob1) and Rex1 (Zfp42); Fig. 2b and Extended Data Fig. 3a) in a manner comparable to those seen in ES cells24. Moderate levels of expression of these pluripotency marker genes were observed on day 3

Viable cells per visual field Oct4-GFP Bright-field

Events

Oct4-GFP

Rearranged DNA

Sorted Oct4-GFP 2 Sorted Oct4-GFP 1

Lymphocytes Fibroblasts

ES cells

a Oct4 SSEA1 Nanog

E-cadherin

ARTICLE RESEARCH Figure 2 | Low-pH-induced Oct4-GFP1 cells

represent pluripotent cells. a, Immunostaining for pluripotent cell markers (red) in day 7 Oct4- GFP1 (green) clusters. DAPI, white. Scale bar, 50 mm. b, qPCR analysis of pluripotency marker genes. From left to right, mouse ES cells; parental CD451 cells; low-pH-induced Oct4-GFP1 cells on day 3; low-pH-induced Oct4-GFP1 cells on day 7. n 5 3; error bars show average 6 s.d. c, DNA methylation study by bisulphite sequencing. Filled and open circles indicate methylated and non- methlylated CpG, respectively. d, Immunostaining analysis of in vitro differentiation capacity of day 7 Oct4-GFP1 cells. Ectoderm: the neural markers Sox1/Tuj1 (100%, n 5 8) and N-cadherin (100%, n 5 5). Mesoderm: smooth muscle actin (50%,

n 5 6) and brachyury (40%, n 5 5). Endoderm: Sox17/E-cadherin (67%, n 5 6) and Foxa2/Pdgfra (67%, n 5 6). Scale bar, 50 mm. e, Teratoma formation assay of day 7 clusters of Oct4-GFP1 cells. Haematoxylin and eosin staining showed keratinized epidermis (ectoderm), skeletal muscle (mesoderm) and intestinal villi (endoderm), whereas immunostaining showed expression of Tuj1 (neurons), smooth muscle actin and a-fetoprotein. Scale bar, 100 mm. f–i, Dissociation culture of ES cells and STAP cells (additional 7 days from day 7; f, g) on gelatin-coated dishes. Top, bright-field; bottom, alkaline phosphatase (AP) staining. Partially dissociated STAP cells slowly generated small colonies (i), whereas dissociated STAP cells did not, even in the presence of the ROCK inhibitor (g, h), which allows dissociation culture of EpiSCs29.

20 15 10

5 0

ES cells Oct4 promoter

Nanog promoter

Ectoderm

Oct4-GFP+ cells Oct4 promoter

Nanog promoter

CD45+ cells Cultured CD45+ cells

Oct4 promoter

Nanog promoter

CD45+ cells

Oct4-GFP+ cells d3 d7

ES cells

Ectoderm

Sox1/Tuj

Mesoderm Endoderm

Mesoderm

α-smooth muscle actin

Brachyury Sox17/E-cadherin

N-cadherin/Sox1

α-smooth muscle actin

Foxa2/Pdgfrα

βIII-tubulin

STAP cells

ES cells Single dissociation

Single dissociation

g STAP cells Single dissociation

Single dissociation

i STAP cells Partial dissociation

Partial dissociation

Day 0

Day 3

Day 0

Day 7

Single dissociation +Y-27632

Day 0

Day 7

AP Day 7

(Fig. 2b and Extended Data Fig. 3b). Notably, the Oct4-GFP1 cells on day 3, but not on day 7, expressed early haematopoietic marker genes such as Flk1 (also called Kdr) and Tal1 (Extended Data Fig. 3c), indicating that Oct4-GFP1 cells on day 3, as judged by their expression pattern at the population level, were still in a dynamic process of conversion.

On day 7, unlike CD451 cells and like ES cells, low-pH-induced Oct4- GFP1 cells displayed extensive demethylation at the Oct4 and Nanog promoter areas (Fig. 2c), indicating that these cells underwent a substantial reprogramming of epigenetic status in these key genes for pluripotency.

In vitro differentiation assays25–27 demonstrated that low-pH-induced Oct4-GFP1 cells gave rise to three-germ-layer derivatives (Fig. 2d) as well as visceral endoderm-like epithelium (Extended Data Fig. 3d). When grafted into mice, low-pH-induced Oct4-GFP1 cell clusters formed teratomas (40%, n 5 20) (Fig. 2e and Extended Data Fig. 4a–c) but no teratocarcinomas that persistently contained Oct4-GFP1 cells (n 5 50). Because some cellular variation was observed in the signal levels of Oct4-GFP within the clusters, we sorted GFP-strong cells (a major popu- lation) and GFP-dim cells (a minor population) by FACS on day 7 and separately injected them into mice. In this case, only GFP-strong cells formed teratomas (Extended Data Fig. 4d). In quantitative polymerase chain reaction (qPCR) analysis, the GFP-strong population expressed

pluripotency marker genes but not early lineage-specific marker genes, whereas the GFP-dim cells showed substantial expression of some early lineage-specific marker genes (Flk1, Gata2, Gata4, Pax6 and Sox17; Extended Data Fig. 4e) but not Nanog and Rex1. These observations indicate that three-germ-layer derivatives were generated from the GFP- strong cells expressing pluripotency marker genes, rather than from GFP-dim cells that seem to contain partially reprogrammed cells.

Collectively, these findings show that the differentiation state of a committed somatic cell lineage can be converted into a state of pluri- potency by strong stimuli given externally. Hereafter, we refer to the fate conversion from somatic cells into pluripotent cells by strong external stimuli such as low pH as ‘stimulus-triggered acquisition of pluripotency’ (STAP) and the resultant cells as STAP cells. Under their establishment conditions, these STAP cells were rarely proliferative (Extended Data Figs 2d and 5a, b). Comparative genomic hybridiza- tion array analysis of STAP cells indicated no major global changes in chromosome number (Extended Data Fig. 5c).

STAP cells compared to ES cells

STAP cells, unlike mouse ES cells, showed a limited capacity for self- renewal in the LIF-containing medium and did not efficiently form

30 JANUARY 2014 | VOL 505 | NATURE | 643

Oct4 Nanog Sox2 Ecat1 Esg1 Dax1

Oct4 promoter

Nanog promoter

Endoderm

α-fetoprotein

Normalized to Gapdh

Oct4-GFP +DAPI Marker

RESEARCH ARTICLE

colonies in dissociation culture (Fig. 2f, g), even in the presence of the ROCK inhibitor Y-27632, which suppresses dissociation-induced apoptosis28,29 (Fig. 2h). Also, even under high-density culture condi- tions after partial dissociation (Fig. 2i), STAP cell numbers started to decline substantially after two passages. Furthermore, expression of the ES cell marker protein Esrrb was low in STAP cells (Extended Data Fig. 5d, e). In general, female ES cells do not show X-chromosomal inactivation30 and contain no H3K27me3-dense foci (indicative of inac- tivated X chromosomes), unlike female CD451 cells and EpiSCs. In contrast, H3K27me3-dense foci were found in ,40% of female STAP cells strongly positive for Oct4-GFP (Extended Data Fig. 5f, g).

STAP cells were also dissimilar to mouse EpiSCs, another category of pluripotent stem cell21,22,29,31, and were positive for Klf4 and negative for the epithelial tight junction markers claudin 7 and ZO-1 (Extended Data Fig. 5d, e).

STAP cells from other tissue sources

We next performed similar conversion experiments with somatic cells collected from brain, skin, muscle, fat, bone marrow, lung and liver tissues of 1-week-old Oct4-gfp mice. Although conversion efficacy varied, the low-pH-triggered generation of Oct4-GFP1 cells was observed in day 7 culture of all tissues examined (Fig. 3a and Extended Data Fig. 6a–c), including mesenchymal cells of adipose tissues (Fig. 3a–c) and neonatal cardiac cells that were negatively sorted for CD45 by FACS (Fig. 3d–g; see Extended Data Fig. 6d for suppression of cardiac genes such as Nkx2-5 and cardiac actin).

Chimaera formation and germline transmission in mice

We next performed a blastocyst injection assay with STAP cells that were generated from CD451 cells of neonatal mice constitutively express- ing GFP (this C57BL/6 line with cag-gfp transgenes is referred to here- after as B6GFP). We injected STAP cell clusters en bloc that were manually cut into small pieces using a microknife (Fig. 4a). A high-to-moderate contribution of GFP-expressing cells was seen in the chimaeric embryos (Fig. 4b and Extended Data Fig. 7a). These chimaeric mice were born

at a substantial rate and all developed normally (Fig. 4c and Extended Data Fig. 7b).

CD451 cell-derived STAP cells contributed to all tissues examined (Fig. 4d). Furthermore, offspring derived from STAP cells were born to the chimaeric mice (Fig. 4e and Extended Data Fig. 7c), demon- strating their germline transmission, which is a strict criterion for pluri- potency as well as genetic and epigenetic normality32,33. Furthermore, in a tetraploid (4N) complementation assay, which is considered to be the most rigorous test for developmental potency34,35 (Fig. 4a, bottom), CD451 cell-derived STAP cells (from F1 mice of B6GFP 3 129/Sv or DBA/2) generated all-GFP1 embryos on embryonic day (E)10.5 (Fig. 4f, Extended Data Fig. 7d and Supplementary Video 3), demon- strating that STAP cells alone are sufficient to construct an entire embry- onic structure. Thus, STAP cells have the developmental capacity to differentiate into all somatic-cell lineages as well as germ-cell lineages in vivo.

Expandable pluripotent cell lines from STAP cells

STAP cells have a limited self-renewal capacity under the conditions used for establishment (Fig. 2g and Extended Data Figs 2e and 5a). However, in the context of the embryonic environment, a small frag- ment of a STAP cell cluster could grow even into a whole embryo (Fig. 4f). With this in mind, we next examined whether STAP cells have the potential to generate expandable pluripotent cell lines in vitro under certain conditions.

STAP cells could not be efficiently maintained for additional pas- sages in conventional LIF1FBS-containing medium or 2i medium20 (most STAP cells died in 2i medium within 7 days; Extended Data Fig. 8a). Notably, an adrenocorticotropic hormone (ACTH)1LIF- containing medium (hereafter called ACTH medium) known to facil- itate clonal expansion of ES cells36 supported outgrowth of STAP cell colonies. When cultured in this medium on a MEF feeder or gelatin, a portion of STAP cell clusters started to grow (Fig. 5a, bottom; such outgrowth was typically found in 10–20% of wells in single cluster culture using 96-well plates and in .75% when 12 clusters were plated

50 40 30 20 10

Adipose-derived mesenchymal cells

***

Control

Low-pH-treated cells

Oct4 Nanog Sox2 Klf4 Rex1

***

CD45+ cells Bone marrow (monocytes) Brain

Lung Muscle Adipose (mesenchyme) Fibroblasts Liver Chondrocytes

d8 7 6 5 4 3 2 1 0

45 40 35 30 25 20 15 10

5 0

d0 d3 d7

Control

Low pH

Oct4/DAPI

Oct4-GFP

Nanog/DAPI SSEA1/DAPI

Cardiomyocyte STAP d7 Oct4-GFP

7 6 Cardiomyocytes 5Oct4 43 Nanog 2Rex1 1 0

Figure 3 | STAP cell conversion from a variety of cells by low-pH treatment. a, Percentage of Oct4-GFP1 cells in day 7 culture of low-pH-treated cells from different origins (1 3 105 cells per ml 3 3 ml). The number of surviving cells on day 7 compared to the plating cell number was 20–30%, except for lung, muscle and adipose cells, for which surviving cells were ,10% (n53,

average 6 s.d.). b, Oct4-GFP1 cell clusters were induced by low-pH treatment from adipose-tissue-derived mesenchymal cells on day 7. Scale bar, 100 mm. c, Expression of pluripotent cell markers in day 7 clusters of low-pH-treated

644 | NATURE | VOL 505 | 30 JANUARY 2014

adipose-tissue-derived mesenchymal cells. Scale bar, 50 mm. d, Expression of pluripotency marker genes in STAP cells derived from various tissues. Gene expressions were normalized by Gapdh (n 5 3, average 6 s.d.). Asterisk indicates adipose tissue-derived mesenchymal cells. e, Quantification of Oct4-GFP1 cells in culture of low-pH-treated neonatal cardiac muscle cells. ***P , 0.001; Tukey’s test (n 5 3). f, Generation of Oct4-GFP1 cell clusters (d7) from CD452 cardiac muscle cells. g, qPCR analysis of pluripotency marker genes in STAP cells from CD452 cardiac muscle cells.

ES Bone marrow

Lung Muscle

Brain Condrocyte

Adipose* Fibroblast

Liver CD45

BM STAP Lung STAP Muscle STAP

Brain STAP Condro STAP

Adipose STAP Fibro STAP

Liver STAP CD45 STAP

Control STAP d3

STAP d7

ES E7.5

Normalized to Gapdh

Oct4-GFP

Percentage of Oct4-GFP+ cells (d7)

Bright-field

Percentage of Oct4-GFP+ cells

ARTICLE RESEARCH

2N blastocyst

4N blastocyst

Bright-field

cag-gfp

Chimaera pup 1 Chimaera pup 2 Chimaera pup 3 Chimaera pup 4 Chimaera pup 5

cag-gfp

STAP stem cells expressed protein and RNA markers for pluripo- tent cells (Fig. 5d, e), showed low DNA methylation levels at the Oct4 and Nanog loci (Extended Data Fig. 8d), and had a nuclear fine struc- ture similar to that of ES cells (Extended Data Fig. 8e; few electron- dense areas corresponding to heterochromatin). In differentiation culture25–27, STAP stem cells generated ectodermal, mesodermal and endodermal derivatives in vitro (Fig. 5f–h and Extended Data Fig. 8f, g), including beating cardiac muscles (Supplementary Video 4), and formed teratomas in vivo (Fig.5i and Extended Data Fig. 8h; no teratocarci- nomas, n 5 40). After blastocyst injection, STAP stem cells efficiently contributed to chimaeric mice (Fig. 5j), in which germline transmis- sion was seen (Extended Data Fig. 8i). Even in tetraploid complemen- tation assays, injected STAP stem cells could generate mice capable of growing to adults and producing offspring (Fig. 5k, l; in all eight inde- pendent lines, Extended Data Fig. 8j).

In addition to their expandability, we noticed at least two other differences between STAP stem cells and parental STAP cells. First, the expression of the ES cell marker protein Esrrb, which was unde- tectable in STAP cells (Extended Data Fig. 5d, e), was clearly seen in STAP stem cells (Fig. 5e). Second, the presence of H3K27me3 foci, which was found in a substantial proportion of female STAP cells, was no longer observed in STAP stem cells (Extended Data Figs 5f and 8k). Thus, STAP cells have the potential to give rise to expandable cell lines that exhibit features similar to those of ES cells.

Discussion

This study has revealed that somatic cells latently possess a surprising plasticity. This dynamic plasticity—the ability to become pluripotent cells—emerges when cells are transiently exposed to strong stimuli that they would not normally experience in their living environments.

Low-pH treatment was also used in the ‘autoneuralization’ experi- ment15–17 by Holtfreter in 1947, in which exposure to acidic medium caused tissue-autonomous neural conversion of salamander animal caps in vitro in the absence of Spemann’s organizer signals. Although the mechanism has remained elusive, Holtfreter hypothesized that the strong stimulus releases the animal cap cells from some intrinsic inhib- itory mechanisms that suppress fate conversion or, in his words, they pass through ‘sublethal cytolysis’ (meaning stimulus-evoked lysis of the cell’s inhibitory state)15,37. Although Holtfreter’s study and ours differ in the direction of fate conversion—orthograde differentiation and nuc- lear reprogramming, respectively—these phenomena may share some common aspects, particularly with regard to sublethal stimulus-evoked release from a static (conversion-resisting) state in the cell.

A remaining question is whether cellular reprogramming is initiated specifically by the low-pH treatment or also by some other types of sub- lethal stress such as physical damage, plasma membrane perforation, osmotic pressure shock, growth-factor deprivation, heat shock or high Ca21 exposure. At least some of these stressors, particularly physical damage by rigorous trituration and membrane perforation by strepto- lysin O, induced the generation of Oct4-GFP1 cells from CD451 cells (Extended Data Fig. 9a; see Methods). These findings raise the possi- bility that certain common regulatory modules, lying downstream of these distantly related sublethal stresses, act as a key for releasing somatic cells from the tightly locked epigenetic state of differentiation, leading to a global change in epigenetic regulation. In other words, unknown cellular functions, activated by sublethal stimuli, may set somatic cells free from their current commitment to recover the naive cell state.

Our present finding of an unexpectedly large capacity for radical reprogramming in committed somatic cells raises various important questions. For instance, why, and for what purpose, do somatic cells latently possess this self-driven ability for nuclear reprogramming, which emerges only after sublethal stimulation, and how, then, is this repro- gramming mechanism normally suppressed? Furthermore, why isn’t teratoma (or pluripotent cell mass) formation normally seen in in vivo tissues that may receive strong environmental stress? In our prelim- inary study, experimental reflux oesophagitis locally induced moderate

30 JANUARY 2014 | VOL 505 | NATURE | 645

c 129/Sv x B6GFP STAP chimaeric mice

Chimaera pup 6 Chimaera pup 7 Chimaera pup 8 Chimaera pup 9 Average

Germline transmission

Bright- field

100 90 80 70 60 50 40 30 20 10 0

E10.5

Figure 4 | Chimaeric mouse generation from STAP cells. a, Schematic of chimaeric mouse generation. b, E13.5 chimaera fetuses from 2N blastocytes injected with STAP cells (derived from B6GFP CD451 cells carrying cag-gfp). c, Adult chimaeric mice generated by STAP-cell (B6GFP 3 129/Sv; agouti) injection into blastocysts (ICR strain; albino). Asterisk indicates a highly contributed chimaeric mouse. d, Chimaera contribution analysis. Tissues from nine pups were analysed by FACS. e, Offspring of chimaeric mice derived from STAP cells. Asterisk indicates the same chimaeric mouse shown in

c. f, E10.5 embryo generated in the tetraploid complementation assay with STAP cells (B6GFP 3 129/Sv).

per well). These growing colonies looked similar to those of mouse ES cells and expressed a high level of Oct4-GFP.

After culturing in ACTH medium for 7 days, this growing popu- lation of cells, unlike parental STAP cells, could be passaged as single cells (Fig. 5a, bottom, and Fig. 5b), grow in 2i medium (Extended Data Fig. 8a) and expand exponentially, up to at least 120 days of culture (Fig. 5c; no substantial chromosomal abnormality was seen; Extended Data Fig. 8b, c). Hereafter, we refer to the proliferative cells derived from STAP cells as STAP stem cells.

Brain Skin

Lung Liver

Heart Muscle

Chimaera contribution ratio (%)

RESEARCH ARTICLE

a b

Bright-field Oct4-GFP

c 1060

1040

1020

ES STAP stem cell MEF STAP

0 40 80 120 Number of days

Ectoderm

Chimaeric mice (2N blastocysts)

Mesoderm

cag-gfp

Endoderm

Single-cell dissociation

d0 AP

d3 1

k Chimaeric mice (4N blastocysts)

3 Oct4 2.5 Nanog

Rex1 2 Klf4

1.5 Sox2

fgh

Oct4

Nanog Klf4

Esrrβ

Ectoderm

Rx/Pax6

Mesoderm Endoderm

Troponin-T Sox17/Ecad

cag-gfp

Germline transmission (4N blastocysts)

1 0.5

Klf2 Esrrb Tbx3

0 Klf5

Figure 5 | ES-cell-like stem cells can be derived from STAP cells. a, Growth of STAP stem cells carrying Oct4-gfp. Scale bar, 50 mm. b, Dissociation culture of STAP stem cells to form colonies. Scale bar, 100 mm. c, Robust growth of STAP stem cells in maintenance culture. Similar results were obtained with eight independent lines. In contrast, parental STAP cells decreased in number quickly. d, Immunostaining of STAP stem cells for pluripotency markers (red). Scale bar, 50 mm. e, qPCR analysis of pluripotency marker gene expression. f–h, In vitro differentiation assays into three-germ-layer derivatives.

f, Ectoderm: Rx1/Pax61 (retinal epithelium; 83%, n 5 6). g, Mesoderm: expression of Oct4-GFP but not endogenous Nanog in the mouse

oesophageal mucosa (Extended Data Fig. 9b). Therefore, an intriguing hypothesis for future research is that the progression from initial Oct4 activation to further reprogramming is suppressed by certain inhib- itory mechanisms in vivo.

The question of why and how this self-driven reprogramming is directed towards the pluripotent state is fundamentally important, given that STAP reprogramming takes a remarkably short period, only a few days for substantial expression of pluripotency marker genes, unlike transgene- or chemical-induced iPS cell conversion38. Thus, our results cast new light on the biological meaning of diverse cellular states in multicellular organisms.

METHODS SUMMARY

Tissue collection and low-pH exposure. To isolate haematopoietic cells, spleens were excised from 1-week-old Oct4-gfp C57BL/6 mice, minced by scissors and mechanically dissociated with pasture pipettes. Dissociated cells were collected, re-suspended in DMEM medium and added to the same volume of lympholyte (Cedarlane), then centrifuged at 1,000g for 20 min. CD45-positive cells were sorted by FACS Aria (BD Biosciences), and treated with low-pH HBSS solution (pH 5.7 for 25 min at 37 uC), centrifuged for 5 min to remove supernatant, and plated to non- adhesive culture plates in DMEM/F12 medium supplemented with 1,000 U LIF (Sigma) and B27 (Invitrogen). Although Oct4-GFP1 cells (expressing pluripotency- related protein and gene markers and capable of differentiating into three germ- layer derivatives) were also generated from lymphocytes of young adult mice (for example, 6-week-old) under the same culture conditions, their proportion in culture was reduced by several to ten folds as compared to neonatal lymphocytes when lymphocytes were isolated from 1-month-old mice or older. Live imaging was performed using specially assembled confocal microscope systems with a CO2 incubator39, and CD45 immunoreactivity in live cells was examined as described40. In vivo and in vitro differentiation assay. STAP cells were seeded onto a sheet 3 3 3 3 1 mm, composed of a non-woven mesh of polyglycolic acid fibres and implanted subcutaneously into the dorsal flanks of 4-week-old NOD/SCID mice. To examine in vitro differentiation, STAP cells and STAP stem cells were collected

646 | NATURE | VOL 505 | 30 JANUARY 2014

troponin-T1 (cardiac muscle; 50%, n 5 6). h, Endoderm: Sox171/E-cadherin1 (endodermal progenitors; 67%, n 5 6). Scale bar, 50 mm. i, Teratoma formation assays. Formation of keratinized epidermis (ectoderm; left), cartilage (mesoderm; middle) and bronchial-like epithelium (endoderm; right) is shown. Scale bar, 100 mm. j, Blastocyst injection assays. These pictures of live animals were taken serially (asterisk indicates the same chimaeric pup).

k, l, Tetraploid complementation assay. ‘All-GFP1’ pups were born (k) and germline transmission was observed (l).

at 7 days and subjected to SDIA or SFEBq culture25,26 for neural differentiation and to embryoid body culture for mesodermal and endodermal27 differentiation.

Online Content Any additional Methods, Extended Data display items and Source Data are available in the online version of the paper; references unique to these sections appear only in the online paper.

Received 10 March; accepted 20 December 2013.

1. Gurdon, J. B. The developmental capacity of nuclei taken from intestinal epithelium cells of feeding tadpoles. J. Embryol. Exp. Morphol. 10, 622–640 (1962). 2. Wakayama, T., Perry, A. C., Zuccotti, M., Johnson, K. R. & Yanagimachi, R. Full-term development of mice from enucleated oocytes injected with cumulus cell nuclei.

Nature 394, 369–374 (1998). 3. Takahashi, K. & Yamanaka, S. Induction of pluripotent stem cells from mouse

embryonic and adult fibroblast cultures by defined factors. Cell 126, 663–676

(2006). 4. Thorpe, T. A. History of plant tissue culture. Mol. Biotechnol. 37, 169–180 (2007). 5. Jiang, Y. et al. Pluripotency of mesenchymal stem cells derived from adult marrow.

Nature 418, 41–49 (2002). 6. D’Ippolito, G. et al. Marrow-isolated adult multilineage inducible (MIAMI) cells, a

unique population of postnatal young and old human cells with extensive

expansion and differentiation potential. J. Cell Sci. 117, 2971–2981 (2004). 7. Kucia, M. et al. A population of very small embryonic-like (VSEL) CXCR41SSEA-

11Oct-41 stem cells identified in adult bone marrow. Leukemia 20, 857–869

(2006). 8. Kuroda, Y. et al. Unique multipotent cells in adult human mesenchymal cell

populations. Proc. Natl Acad. Sci. USA 107, 8639–8643 (2010). 9. Obokata, H. et al. The potential of stem cells in adult tissues representative of the

three germ layers. Tissue Eng. Part A 17, 607–615 (2011). 10. Lengner, C. J., Welstead, G. G. & Jaenisch, R. The pluripotency regulator Oct4: a role

in somatic stem cells? Cell Cycle 7, 725–728 (2008). 11. Berg,J.S.&Goodell,M.A.AnargumentagainstaroleforOct4insomaticstem

cells. Cell Stem Cell 1, 359–360 (2007). 12. Hong,H.etal.Suppressionofinducedpluripotentstemcellgenerationbythe

p53-p21 pathway. Nature 460, 1132–1135 (2009). 13. Hanna, J. et al. Direct reprogramming of terminally differentiated mature B

lymphocytes to pluripotency. Cell 133, 250–264 (2008). 14. Ohbo,K.etal.Identificationandcharacterizationofstemcellsinprepubertal

spermatogenesis in mice small star, filled. Dev. Biol. 258, 209–225 (2003). 15. Holtfreter,J.Neuralinductioninexplantswhichhavepassedthroughasublethal

cytolysis. J. Exp. Zool. 106, 197–222 (1947).

ES STAP stem cell

+DAPI Antibodies Passage 1 Day 7 Day 1

Relative expression levels (ES=1.0)

Cell number

ARTICLE

RESEARCH

16. Gerhart, J. Johannes Holtfreter’s contributions to ongoing studies of the organizer. Dev. Dyn. 205, 245–256 (1996).

17. Byrnes, W. M. Ernest Everett Just, Johannes Holtfreter, and the origin of certain concepts in embryo morphogenesis. Mol. Reprod. Dev. 76, 912–921 (2009).

18. Gratama, J. W., Sutherland, D. R. & Keeney, M. Flow cytometric enumeration and immunophenotyping of hematopoietic stem and progenitor cells. Semin. Hematol. 38, 139–147 (2001).

19. Inoue, K. et al. Inefficient reprogramming of the hematopoietic stem cell genome following nuclear transfer. J. Cell Sci. 119, 1985–1991 (2006).

20. Ying, Q. L. et al. The ground state of embryonic stem cell self-renewal. Nature 453, 519–523 (2008).

21. Tesar, P. J. et al. New cell lines from mouse epiblast share defining features with human embryonic stem cells. Nature 448, 196–199 (2007).

22. Brons, I. G. et al. Derivation of pluripotent epiblast stem cells from mammalian embryos. Nature 448, 191–195 (2007).

23. Kuroda,Y.etal.Isolation,cultureandevaluationofmultilineage-differentiating stress-enduring (Muse) cells. Nature Protocols 8, 1391–1415 (2013).

24. Niwa,H.Howispluripotencydeterminedandmaintained?Development134, 635–646 (2007).

25. Watanabe, K. et al. Directed differentiation of telencephalic precursors from embryonic stem cells. Nature Neurosci. 8, 288–296 (2005).

26. Kawasaki,H.etal.InductionofmidbraindopaminergicneuronsfromEScellsby stromal cell-derived inducing activity. Neuron 28, 31–40 (2000).

27. Gouon-Evans,V.etal.BMP-4isrequiredforhepaticspecificationofmouse embryonic stem cell-derived definitive endoderm. Nature Biotechnol. 24, 1402–1411 (2006).

28. Ohgushi,M.&Sasai,Y.Lonelydeathdanceofhumanpluripotentstemcells: ROCKing between metastable cell states. Trends Cell Biol. 21, 274–282 (2011). 29. Ohgushi,M.etal.Molecularpathwayandcellstateresponsiblefordissociation-

induced apoptosis in human embryonic stem cells. Cell Stem Cell 7, 225–239

(2010). 30. Murakami,K.,Araki,K.,Ohtsuka,S.,Wakayama,T.&Niwa,H.Choiceofrandom

rather than imprinted X inactivation in female embryonic stem cell-derived extra-

embryonic cells. Development 138, 197–202 (2011). 31. Pera,M.F.&Tam,P.P.Extrinsicregulationofpluripotentstemcells.Nature465,

713–720 (2010). 32. Surani, M. A. & Barton, S. C. Development of gynogenetic eggs in the mouse:

implications for parthenogenetic embryos. Science 222, 1034–1036 (1983). 33. Wakayama, S. et al. Successful serial recloning in the mouse over multiple

generations. Cell Stem Cell 12, 293–297 (2013).

34. Nagy, A., Rossant, J., Nagy, R., Abramow-Newerly, W. & Roder, J. C. Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc. Natl Acad. Sci. USA 90, 8424–8428 (1993).

35. Eakin, G. S., Hadjantonakis, A. K., Papaioannou, V. E. & Behringer, R. R. Developmental potential and behavior of tetraploid cells in the mouse embryo. Dev. Biol. 288, 150–159 (2005).

36. Ogawa, K., Matsui, H., Ohtsuka, S. & Niwa, H. A novel mechanism for regulating clonal propagation of mouse ES cells. Genes Cells 9, 471–477 (2004).

37. Hurtado, C. & De Robertis, E. M. Neural induction in the absence of organizer in salamander is mediated by MAPK. Dev. Biol. 307, 282–289 (2007).

38. Hou, P. et al. Pluripotent stem cells induced from mouse somatic cells by small-molecule compounds. Science 341, 651–654 (2013).

39. Eiraku, M. et al. Self-organizing optic-cup morphogenesis in three-dimensional culture. Nature 472, 51–56 (2011).

40. Eilken,M.H.,Nishikawa,S.&Schroeder,T.Continuoussingle-cellimaging of blood generation from haemogenic endothelium. Nature 457, 896–900 (2009).

Supplementary Information is available in the online version of the paper.

Acknowledgements We thank S. Nishikawa for discussion and J. D. Ross, N. Takata, M. Eiraku, M. Ohgushi, S. Itoh, S. Yonemura, S. Ohtsuka and K. Kakiguchi for help with experiments and analyses. We thank A. Penvose and K. Westerman for comments on the manuscript. H.O. is grateful to T. Okano, S. Tsuneda and K. Kuroda for support and encouragement. Financial support for this research was provided by Intramural RIKEN Research Budget (H.O., T.W. and Y.S.), a Scientific Research in Priority Areas (20062015) to T.W., the Network Project for Realization of Regenerative Medicine to Y.S., and Department of Anesthesiology, Perioperative and Pain Medicine at Brigham and Women’s Hospital to C.A.V.

Author Contributions H.O. and Y.S. wrote the manuscript. H.O., T.W. and Y.S. performed experiments, and K.K. assisted with H.O.’s transplantation experiments. H.O., T.W., Y.S., H.N. and C.A.V. designed the project. M.P.V. and M.Y. helped with the design and evaluation of the project.

Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of the paper. Correspondence and requests for materials should be addressed to H.O. (obokata@cdb.riken.jp) or C.A.V. (cvacanti@partners.org).

30 JANUARY 2014 | VOL 505 | NATURE | 647

RESEARCH

ARTICLE

METHODS

Animal studies. Research involving animals complied with protocols approved by the Harvard Medical School/Brigham and Women’s Hospital Committee on Animal Care, and the Institutional Committee of Laboratory Animal Experimen- tation of the RIKEN Center for Developmental Biology.

Tissue collection and low-pH treatment. To isolate CD451 haematopoietic cells, spleens were excised from 1-week-old Oct4-gfp mice (unless specified otherwise), minced by scissors and mechanically dissociated with pasture pipettes. Dissociated spleen cells were suspended with PBS and strained through a cell strainer (BD Biosciences). After centrifuge at 1,000 r.p.m. for 5 min, collected cells were re-suspended in DMEM medium and added to the same volume of lympholyte (Cedarlane), then centrifuged at 1,000g for 20 min. The lymphocyte layer was taken out and stained with CD45 antibody (ab25603, Abcam). CD45-positive cells were sorted by FACS Aria (BD Biosciences). After cell sorting, 1 3 106 CD45-positive cells were treated with 500 ml of low-pH HBSS solution (titrated to pH5.7 by HCl) for 25 min at 37 uC, and then centrifuged at 1,000 r.p.m. at room temperature for 5 min. After the supernatant (low-pH solution) was removed, precipitated cells were re-suspended and plated onto non-adhesive culture plates (typically, 1 3 105 cells ml21) in DMEM/F12 medium supplemented with 1,000 U LIF (Sigma) and 2% B27 (Invitrogen). Cell cluster formation was more sensitive to the plating cell density than the percentage of Oct4-GFP1 cells. The number of surviving cells was sensitive to the age of donor mice and was low under the treatment conditions above when adult spleens were used. The addition of LIF during days 2–7 was essential for generating Oct4-GFP1 STAP cell clusters on day 7, as shown in Extended Data Fig. 1f. Even in the absence of LIF, Oct4-GFP1 cells (most of them were dim in signal) appeared transiently during days 2–5 in culture of low-pH-treated CD451 cells, but subsequently disappeared, indicating that there is a LIF-independent early phase, whereas the subsequent phase is LIF-dependent.

Chimaeric mouse generation and analyses. For production of diploid and tetra- ploid chimaeras with STAP cells, diploid embryos were obtained from ICR strain females. Tetraploid embryos were produced by electrofusion of 2-cell embryos. Because trypsin treatment of donor samples turned out to cause low chimaerism, STAP spherical colonies were cut into small pieces using a microknife under the microscope, and small clusters of STAP cells were then injected into day-4.5 blas- tocysts by a large pipette. The next day, the chimaeric blastocysts were transferred into day-2.5 pseudopregnant females. For experiments using STAP cells from CD451 cells without the Oct4-gfp reporter, STAP cell clusters were identified by their characteristic cluster morphology (they are made of very small cells with no strong compaction in the aggregate). When the STAP conversion conditions (low pH) were applied to CD451 lymphocytes, most day-7 clusters that were large and contained more than a few dozen small cells were positive for Oct4 (although the expression level varied). Therefore, we used only well-formed characteristic clus- ters (large ones) for this type of study and cut them by microknife to prepare donor cell clusters in a proper size for glass needle injection. For an estimate of the contribution of these injected cells, we used STAP cells that were generated from CD451 cells of mice constitutively expressing GFP (C57BL/6 line with cag-gfp transgenes; F1 of C57BL/6 and 129/Sv or DBA/2 was used from the viewpoint of heterosis).

Because the number of CD451 cells from a neonatal spleen was small, we mixed spleen cells from male and female mice for STAP cell conversion. To make germ- line transmission more efficient, we intercrossed chimaeras in some experiments.

For the production of diploid and tetraploid chimaeras with STAP stem cells, diploid embryos were obtained from ICR strain females. Tetraploid embryos were produced by electrofusion of 2-cell embryos. STAP stem cells were dissociated into single cells and injected into day-4.5 blastocysts. In the chimaera studies with both STAP cells and STAP stem cells, we did not find tumorigenetic tendencies in their chimaeras or their offspring (up to 18 months).

In vivo differentiation assay. 1 3 107 STAP cells were seeded onto a sheet com- posed of a non-woven mesh of polyglycolic acid fibres (3 3 3 3 1 mm; 200 mm in pore diameter), cultured for 24 h in DMEM 1 10% FBS, and implanted subcu- taneously into the dorsal flanks of 4-week-old mice. In this experiment, to better support tumour formation from slow growing STAP cells by keeping cells in a locally dense manner, we implanted STAP cells with artificial scaffold made of polyglycolic acid fibres. Given the artificial nature of the material, we used NOD/ SCID mice as hosts, to avoid possible enhancement of post-graft inflammation caused by this scaffold even in syngenic mice. STAP stem cells were dissociated into single cells and cell suspension containing 1 3 107 cells was injected into the testis. Six weeks later, the implants were analysed using histochemical techniques. The implants were fixed with 10% formaldehyde, embedded in paraffin, and routinely processed into 4-mm-thick sections. Sections were stained with haema- toxylin and eosin. Endoderm tissues were identified with expression of anti-a- fetoprotein (mouse monoclonal antibody; MAB1368, R&D Systems). Ectodermal tissues were identified with expression of anti-bIII tubulin (mouse monoclonal

antibody; G7121, Promega). Mesodermal tissues were identified with expression of anti-a-smooth muscle actin (rabbit polyclonal; DAKO). In negative controls, the primary antibody was replaced with IgG-negative controls of the same isotype to ensure specificity.

STAP by exposure to other external stimuli. To give a mechanical stress to mature cells, a pasture pipette was heated and then stretched to create thin capillaries with the lumens approximately 50 mm in diameter, and then broken into appropriate lengths. Mature somatic cells were then repeatedly triturated through these pipettes for 20 min, and then cultured for 7 days. To provide a heat shock, cells were heated at 42 uC for 20 min and cultured for 7 days. A nutrition-deprivation stress was pro- vided to mature cells, by culturing the cells in basal culture medium for 3 weeks. High Ca21 concentration stress was provided to mature cells by culturing cells in medium containing 2 mM CaCl2 for 7 days. To give a strong stress by creating pores in cell membranes, cells were treated with 230 ng ml21 streptolysine O (SLO) (S5265, Sigma) for 2 h, then cultured for 7 days. After each treatment, the ratio of Oct4-GFP-positive cells was analysed by FACS.

Bisulphite sequencing. GFP-positive cells in STAP clusters were collected by FACS Aria. Genomic DNA was extracted from STAP cells and analysed. Bisulphite treat- ment of DNA was performed using the CpGenome DNA modification kit (Chemicon, http://www.chemicon.com), following the manufacturer’s instructions. The result- ing modified DNA was amplified by nested PCR using two forward (F) primers and one reverse (R) primer: Oct4 (F1, 59-GTTGTTTTGTTTTGGTTTTGGATA T-39; F2, 59-ATGGGTTGAAATATTGGGTTTATTTA-39; R, 59-CCACCCTCT AACCTTAACCTCTAAC-39). And Nanog (F1, 59-GAGGATGTTTTTTAAGT TTTTTTT-39; F2, 59-AATGTTTATGGTGGATTTTGTAGGT-39; R, 59-CCCA CACTCATATCAATATAATAAC-39). PCR was done using TaKaKa Ex Taq Hot Start Version (RR030A). DNA sequencing was performed using a M13 primer at the Genome Resource and Analysis Unit, RIKEN CDB. Immunohistochemistry. Cultured cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100/PBS before blocking with 1% BSA solution. Cells were incubated with the following primary antibodies: anti-Oct4 (Santa Cruz Biotechnology; C-10), anti-Nanog (eBioscience; MLC-51), anti-SSEA-1 (Millipore; MC480), anti-E-cadherin (Abcam), anti-ZO-1 (Santa Cruz Biotech- nology; c1607), anti-claudin7 (Abcam), anti-Klf4 (R&D Systems), anti-Esrrb (R&D Systems), anti-H3K27me3 (Millipore), anti-BrdU (BD Bioscience) and anti-Ki67 (BD Pharmingen). After overnight incubation, cells were incubated with secondary antibodies: goat anti-mouse or -rabbit coupled to Alexa-488 or -594 (Invitrogen). Cell nuclei were visualized with DAPI (Sigma). Slides were mounted with a SlowFade Gold antifade reagent (Invitrogen).

Fluorescence-activated cell sorting and flow cytometry. Cells were prepared according to standard protocols and suspended in 0.1% BSA/PBS on ice before FACS. Propidium iodide (BD Biosciences) was used to exclude dead cells. In nega- tive controls, the primary antibody was replaced with IgG-negative controls of the same isotype to ensure specificity. Cells were sorted on a BD FACSAria SORP and analysed on a BD LSRII with BD FACS Diva Software (BD Biosciences). For hae- matopoietic fraction sorting, antibodies against T-cell marker (anti-CD90; eBioscience), B-cell marker (anti-CD19; Abcam) and haematopoietic progenitor marker (anti- CD34; Abcam) were used.

RNA preparation and RT–PCR analysis. RNA was isolated with the RNeasy Micro kit (Qiagen). Reverse transcription was performed with the SuperScript III first strand synthesis kit (Invitrogen). Power SYBR Green Mix (Roche Diagnostics) was used for amplification, and samples were run on a Lightcycler-II Instrument (Roche Diagnostics). The primer sets for each gene are listed in Supplementary Table 1.

In vitro differentiation assays. For mesodermal differentiation assay, STAP cells were collected at 7 days, and Oct4-GFP-positive cells were collected by cell sorter and subjected to culture in DMEM supplemented with 20% FBS. Medium was exchanged every 3 days. After 7–14 days, muscle cells were stained with an anti-a- smooth muscle actin antibody (DAKO).

For neural lineage differentiation assay, STAP cells were collected at 7 days and subjected to SDIA or SFEBq culture. For SDIA culture, collected STAP cell clusters were plated on PA6 cell feeder as described previously26. For SFEBq culture, STAP cell clusters (one per well; non-cell-adhesive 96-well plate, PrimeSurface V-bottom, Sumitomo Bakelite) were plated and cultured in suspension as described previously36.

For endodermal differentiation, STAP cells were collected at 7 days and sub- jected to suspension culture with inducers in 96-well plates27. TCR-b chain gene rearrangement analysis. Genomic DNA was extracted from STAP cells and tail tips from chimaeric mice generated with STAP cells derived from CD451 cells. PCR was performed with 50 ng DNA using the following primers (Db2: 59-GCACCTGTGGGGAAGAAACT-39 and Jb2.6: 59-TGAGAGCTGTCT CCTACTATCGATT-39) that amplify the regions of the (D)J recombination. The PCR products were subjected to gel electrophoresis in Tris-acetate-EDTA buffer with 1.6% agarose and visualized by staining with ethidium bromide. PCR bands

ARTICLE

RESEARCH

from STAP cells were subjected to sequencing analysis and identified as rear- ranged genomic fragments of the (D)J recombination. EdU uptake assay and apoptosis analysis. At various phases in STAP cell culture (days 0–2, 2–7, 7–14), EdU was added to the culture medium (final concentration: 10mM) and EdU uptake was analysed by FACS. This assay was performed according to the manufacturer’s protocol with the Click-iT EdU Flow cytometry assay kit (Invitrogen).

Apoptosis analysis was performed with flow cytometry using Annexin-V (Bio- vision) and propidium iodide. Annexin-V analysis by FACS on day 14 showed that most Oct4-GFP1 cells were positive for this apoptotic marker; indeed, the number of surviving cells declined thereafter.

Soft agar assay. Sorted STAP cells (Oct4-GFP-strong or -dim) and control mouse ES cells (1,000 cells per well of 96-well plate) were plated into soft ager medium (0.4% agarose) in LIF-B27 medium. After 7 days of culture, cells were dissociated and their anchorage-independent growth was quantified by fluorescent measure- ment with the cytoselect 96-well cell transformation assay kit (Cell Biolabs) accord- ing to the manufacturer’s protocol.

Comparative genomic hybridization (CGH) array analysis. Genomic DNA was extracted from STAP (male) and CD45-positive cells (male) by the Gene JET Genomic DNA purification kit (Thermo Scientific). Using CGH array (Agilent), the normality of chromosomes derived from STAP was compared with that of CD45-positive cells whose chromosomal normality was confirmed by a separate experiment. CGH array and data analysis were performed at TAKARA Bio. Electron microscopy. For electron microscopic analysis, dissociated cells were fixed in 2.5% glutaraldehyde and 2% formaldehyde in 0.1 M cacodylate buffer (pH 7.2) and then processed for thin sectioning and transmission electron microscopy. Live cell imaging. All live-cell imaging was performed with LCV110-CSUW1 (Olympus). For live-cell imaging of ‘in culture CD45 antibody staining’, CD451 cells treated with low pH were plated in culture medium containing 20 ng ml21 of fluorescent-labelled CD45 antibody (eBioscience)40.

RNA-sequencing and ChIP sequencing analyses. For RNA sequencing of cell lines, total RNA was extracted from cells by the RNasy mini kit (Qiagen). RNA-seq libraries were prepared from 1 mg total RNAs following the protocol of the TruSeq RNA Sample Prep kit (Illumina) and subjected to the deep sequencing analysis with Illumina Hi-Seq1500. Cluster tree diagram of various cell types was obtained from hierarchical clustering of global expression profiles (log2 FPKM of all tran- scripts; FPKM, fragments per kilobase of transcript per million mapped reads). Complete linkage method applied to 1 2 r (r 5 Pearson’s correlation between profiles) was used for generating the tree and 1,000 cycles of bootstrap resampling were carried out to obtain statistical confidence score in per cent units (also called AU P values).

ChIP-seq libraries were prepared from 20 ng input DNAs, 1 ng H3K4me3 ChIP DNAs, or 5 ng H3K27me3 ChIP DNAs using the KAPA Library Preparation kit (KAPA Biosystems). TruSeq adaptors were prepared in-house by annealing a TruSeq universal oligonucleotide and each of index oligonucleotides (59-AATGATACG GCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATC T-39, and 59-GATCGGAAGAGCACACGTCTGAACTCCAGTCACXXXXXXA TCTCGTATGCCGTCTTCTGCTTG-39; where X represents index sequences).

Chromatin immunoprecipitation was performed as follows. Cells were fixed in PBS(-) containing 1% formaldehyde for 10 min at room temperature. Glycine was added to a final concentration of 0.25 M to stop the fixation. After washing the cells twice in ice-cold PBS(-), cells were further washed in LB1 (50 mM HEPES- KOH pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 0.25% Triton X-100) and LB2 (10 mM Tris-HCl pH 8.0, 200 mM NaCl, 1 mM EDTA, 0.5 mM EGTA). Cells were then re-suspended in lysis buffer (50 mM Tris-HCl pH 8.0, 10 mM EDTA, 1% SDS). Lysates were prepared by sonication using Covaris S220 in a mini tube at duty cycle 5 5%, PIP 5 70, cycles per burst 5 200, and the

treatment time of 20 min. Lysates from 2 3 106 cells were diluted in ChIP dilution buffer (16.7 mM Tris-HCl pH 8.0, 167 mM NaCl, 1.2 mM EDTA, 1.1% Triton X-100, 0.01% SDS). ChIP was performed using sheep anti-mouse IgG beads (Invitrogen) or protein A beads (Invitrogen) coupled with anti-histone H3K4me3 antibody (Wako, catalogue no. 307-34813) or anti-histone H3K27me3 antibody (CST, cata- logue no. 9733), respectively. After 4–6 h of incubation in a rotator at 4 uC, beads were washed five times in low-salt wash buffer (20 mM Tris HCl pH 8.0, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS), and three times in high-salt wash buffer (20 mM Tris-HCl pH 8.0, 500 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS). Target chromatin was eluted off the beads in elution buffer (10 mM Tris-HCl pH 8.0, 300 mM NaCl, 5 mM EDTA, 1% SDS) at room temperature for 20 min. Crosslink was reversed at 65 uC, and then samples were treated with RNaseA and proteinase K. The prepared DNA samples were purified by phenol-chloroform extraction followed by ethanol precipitation and dissolved in TE buffer.

STAP stem-cell conversion culture. For establishment of STAP stem-cell lines, STAP cell clusters were transferred to ACTH-containing medium36 on MEF feeder cells (several clusters, up to a dozen clusters, per well of 96-well plates). Four to seven days later, the cells were subjected to the first passage using a conventional trypsin method, and suspended cells were plated in ES maintain medium contain- ing 20% FBS. Subsequent passaging was performed at a split ratio of 1:10 every second day before they reached subconfluency. We tested the following three dif- ferent genetic backgrounds of mice for STAP stem-cell establishment from STAP cell clusters, and observed reproducible data of establishment: C57BL/6 carrying Oct4-gfp (29 of 29), 129/Sv carrying Rosa26-gfp (2 of 2) and 129/Sv 3 C57BL/6 carrying cag-gfp (12 of 16). STAP stem cells with all these genetic backgrounds showed chimaera-forming activity.

For clonal analysis of STAP stem cells, single STAP stem cells were manually picked by a thin-glass pipette, and plated into 96-well plates at one cell per well. The clonal colonies were cultured in ES medium containing 20% FBS, and expanded for subsequent experiments.

Karyotype analysis. Karyotype analysis was performed by Multicolor FISH ana- lysis (M-FISH). Subconfluent STAP stem cells were arrested in metaphase by col- cemid (final concentration 0.270 mg ml21) to the culture medium for 2.5 h at 37 uC in 5% CO2. Cells were washed with PBS, treated with trypsin and EDTA (EDTA), re-suspended into cell medium and centrifuged for 5 min at 1,200 r.p.m. To the cell pellet in 3 ml of PBS, 7 ml of a pre-warmed hypotonic 0.0375 M KC1 solution was added. Cells were incubated for 20 min at 37 uC. Cells were centrifuged for 5 min at 1,200 r.p.m. and the pellet was re-suspended in 3–5 ml of 0.0375 M KC1 solution. The cells were fixed with methanol/acetic acid (3:1; vol/vol) by gently pipetting. Fixation was performed four times before spreading the cells on glass slides. For the FISH procedure, mouse chromosome-specific painting probes were combinatorially labelled using seven different fluorochromes and hybridized as previously described41. For each cell line, 9–15 metaphase spreads were acquired by using a Leica DM RXA RF8 epifluorescence microscope (Leica Mikrosysteme GmbH) equipped with a Sensys CCD camera (Photometrics). Camera and microscope were controlled by the Leica Q-FISH software (Leica Microsystems). Metaphase spreads were processed on the basis of the Leica MCK software and presented as multicolour karyograms.

Q-band analysis was performed at Chromocentre (Japan). After quinacrin staining, 20 cells from each sample were randomly selected and the normality of chromosomes was analysed. Five different independent lines of STAP stem cells showed no chromosomal abnormalities in Q-band analysis after .10 passages.

41.

Jentsch, I., Geigl, J., Klein, C. A. & Speicher, M. R. Seven-fluorochrome mouse M-FISH for high-resolution analysis of interchromosomal rearrangements. Cytogenet. Genome Res. 103, 84–88 (2003).

RESEARCH

ARTICLE

Extended Data Figure 1 | Conversion of haematopoietic cells into Oct4-GFP1 cells by a low-pH exposure. a, Optimization of pH conditions for Oct4-GFP induction. Five days after CD45-positive cells were exposed to acidic solution treatment at different pH, Oct4-GFP expression was analysed by FACS (n 5 3, average 6 s.d.). b, Gating strategy for Oct4-GFP1 cell sorting. Top: representative results 7 days after the stress treatment. Bottom: non- treated control. P3 populations were sorted and counted as Oct4-GFP1 cells for all experiments. c, Controls for FACS analysis. In Oct4-GFP1 cell analysis, the grey and white histograms indicate the negative control (non-stress-treated Oct4-gfp haematopoietic cells) and the positive control (Oct4-gfp ES cells), respectively. Also, the green histograms indicate non-treated cells (left) and stress-treated cells at day 7 (right). In CD451 cell analysis, the grey and white histograms indicate the negative (isotype) and positive controls, respectively. The red histograms indicate non-stress-treated cells (left) and stress-treated cells at day 7 (right). d, Oct4-GFP1 cell generation from various subpopulations of CD451 cells. Seven days after the stress treatment, Oct4-GFP expression was analysed by FACS (n 5 3, average 6 s.d.). Among total CD451 fraction and its subfractions of CD191, CD901, CD341 and CD342 cells, the efficacy of CD341 cells was significantly lower than the others. P , 0.05 by the Newman–Keuls test and P , 0.01 by one-way ANOVA. e, Comparison of

culture conditions for low-pH-induced conversion. Stress-treated cells were cultured in various media. The number of Oct4-GFP-expressing clusters was counted at day 14 (n 5 3, average 6 s.d.). ***P , 0.001 (B271LIF versus all other groups); Tukey’s test. In the case of 3i medium, although the clusters appeared at a moderate efficiency, they appeared late and grew slowly. ACTH, ACTH-containing ES medium; ES1LIF1FBS, 20% FBS1LIF-containing ES culture medium; B27, DMEM/F12 medium containing 2% B27; B271LIF, DMEM/F12 medium containing 2% B271LIF; EpiSC, EpiSC culture medium containing Fgf21activin. f, Signalling factor dependency of STAP cell generation. Growth factors that are conventionally used for pluripotent cell culture such as LIF, activin, Bmp4 and Fgf2 were added to basal culture medium (B27-supplemented DMEM/F12) in different culture phases (days 0–7, 2–7 and 4–7), and Oct4-GFP expression was analysed by FACS at day 7 (n 5 3, average 6 s.d.). g, h, Time course of apoptosis after the low-pH exposure. Stress-treated cells and non-stress-treated control cells were stained with CD45, annexin-V and propidium iodide at day 0 (immediately after stress treatment), day 3 and day 7. g, Blue bars, GFP1CD452; orange bars, GFP2CD451. Percentages in total cells included propidium-iodide-positive cells.

h, Annexin-V-positive cells in these cell populations were analysed by FACS.

ARTICLE

RESEARCH

Extended Data Figure 2 | Phenotypic change during STAP cell conversion. a, Oct4 protein expression in STAP cells was detected by immunostaining at day 2 (left) and day 7 (right). b, Live cell imaging of STAP conversion (grey, CD45 antibody; green, Oct4-GFP). See Methods for experimental details to monitor live CD45 immunostaining. c, Immunostaining of a parental CD451 cell (left) and an Oct4-GFP1 cell (right). Scale bar, 10 mm. d, EdU uptake assay (n 5 3, average 6 s.d.). e, Schematic of Tcrb gene rearrangement.

f, T-cell-derived STAP cells. Scale bar, 100 mm. g, Genomic PCR analysis of (D)J recombination at the Tcrb gene of T-cell-derived STAP cells. G.L. is the size of the non-rearranged germline type, whereas the smaller ladders correspond to the alternative rearrangements of J exons (confirmed by sequencing). Negative controls (ES cells), positive controls (lymphocytes) and T-cell-derived STAP (two independent preparations on d7) are indicated.

RESEARCH

ARTICLE

Extended Data Figure 3 | Gene expression analyses during STAP conversionandendodermdifferentiationassay. a,Expressionof pluripotency marker genes in STAP cells derived from T cells (n 5 3, average 6 s.d.). b, Expression of pluripotency marker genes in STAP cells. In this experiment, Oct4-GFP1 cells seen in live cell imaging (Extended Data

Fig. 2b) were analysed to confirm their conversion into STAP cells (n 5 3, average6s.d.).c,HaematopoieticmarkerexpressionduringSTAPconversion from CD451 cells (n 5 3, average 6 s.d.). d, Formation of visceral endoderm-like surface epithelium in differentiating STAP cluster on day 2 (left) and day 8 (right). Scale bars, 50 mm.

ARTICLE

RESEARCH

Extended Data Figure 4 | Teratoma formation assay and characterization of Oct4-GFP-dim cells. a–c, Teratomas formed from STAP cell clusters included neuroepithelium (a), striated muscle (b) and pancreas (c; right, high-magnification view showing a typical acinar morphology and ductal structures). Scale bars, 100 mm. d, Teratoma-forming ability of Oct4-GFP1 and Oct4-GFP-dim cells (isolated by FACS, top). Oct4-GFP1 cells, but not

Oct4-GFP-dim cells, efficiently formed teratomas (table at the bottom). However, because STAP cells were dissociation-intolerant, the teratoma- forming efficiency of dissociated Oct4-GFP1 cells was lower than that of non-dissociated STAP cell clusters. e, Gene expression of Oct4-GFP1 and Oct4-GFP-dim cells (n 5 3, the average 6 s.d.). Haematopoietic marker gene expression (left) and early lineage marker gene expression (right) are shown.

RESEARCH

ARTICLE

Extended Data Figure 5 | In vitro characterization of STAP cells. a, Immunostaining for Ki67 and BrdU. STAP cell clusters (top) and ES cell colonies (bottom) are shown. For BrdU uptake, BrdU was added into each culture medium (10 mM) for 12 h until fixation. Scale bar, 100 mm. b, Transformation assay by soft agar culture. Neither Oct4-GFP1 nor Oct4-GFP-dim cells showed colony formation in soft agar, whereas ES cells and STAP stem cells showed anchorage-independent growth in the same LIF-B27 medium. Scale bar, 100 mm. Proliferated cells were lysed and the amount of DNA in each well was estimated by chemical luminescence (graph). n 5 3 , average 6 s.d. c, No substantial change in chromosome number was seen with STAP cells in the CGH array. Genomic DNA derived from CD451 cells (male) was used as reference DNA. The spikes (for example, those seen in the X chromosome) were nonspecific and also found in the data of these parental

CD451 cells when the manufacturer’s control DNA was used as a reference. d, qPCR analysis for pluripotency markers that highly express in ES cells, but not in EpiSCs. Average 6 s.d. e, Immunostaining of markers for mouse EpiSC and ES cells. Scale bar, 100 mm. f, g, H3K27me31 foci in female cells, which are indicative of X-chromosomal inactivation. These foci were not observed in male cells. Scale bar, 10 mm. In the case of female STAP cells, ,40% of cells retained H3K27me31 foci (g). **P , 0.001; Tukey’s test. n 5 3, average 6 s.d. Although nuclear staining looked to be higher in STAP cells with H3K27me31 foci (f), this appeared to be caused by some optical artefacts scattering from the strong focal signal. h, qPCR analysis for the tight junction markers Zo-1 and claudin 7, which were highly expressed in EpiSCs, but not in ES cells or STAP cells. **P , 0.01; ns, not significant; Tukey’s test; n 5 3, average 6 s.d.

ARTICLE

RESEARCH

Extended Data Figure 6 | Conversion of somatic tissue cells into STAP cells. a, Alkaline phosphatase expression of STAP cells derived from adipose-derived mesenchymal cells. Scale bar, 100 mm. b, E-cadherin expression of STAP cells derived from adipose-derived mesenchymal cells. Scale bar, 50 mm. c, FACS sorting of dissociated neonatal cardiac muscle cells by removing CD451 cells. d, Cardiomyocyte marker gene expression during STAP conversion from cardiomyocytes (n 5 3, average 6 s.d.).

RESEARCH

ARTICLE

Extended Data Figure 7 | Generation chimaeras with STAP cells. a, 2N chimaeras generated with STAP cells derived from Oct4-gfp C57BL/6 mice (left) and 129/Sv 3 C57BL/6 F1 mice (right). b, Generation of chimaeric mice from STAP cells by cluster injection. STAP cells used in the experiments above were generated from CD451 lymphocytes of multiple neonatal spleens

(male and female tissues were mixed). *All fetuses were collected at 13.5 d.p.c. to 15.5 d.p.c. and the contribution rate of STAP cells into each organ was examined by FACS. **The contribution of STAP cells into each chimaera

was scored as high (.50% of the coat colour of GFP expression). ***B6GFP: C57BL/6 mouse carrying cag-gfp. c, Production of offspring from STAP cells via germline transmission. Chimaeras generated with 129/Sv 3 B6GFP STAP cells (obtained from the experiments shown in b) were used for germline transmission study. d, 4N embryos at E9.5 generated with STAP cells derived from F1 GFP mice (B6GFP and DBA/2 or 129/Sv). B6GFP, C57BL/6 mouse carrying cag-gfp.

ARTICLE

RESEARCH

Extended Data Figure 8 | Molecular and cellular characterization of STAP stem cells. a, Compatibility of 2i conditions with STAP stem-cell derivation from STAP cells and STAP stem-cell maintenance. STAP stem cells could not be established directly from STAP cells in 2i 1 LIF medium (top). However, once established in ACTH medium, STAP stem cells were able to survive and expand in 2i 1 LIF medium. Scale bar, 100 mm. b, Q-band analysis (n 5 4; all cell lines showed the normal karyotype). c, Multicolour FISH analysis (n 5 8; all cell lines showed the normal karyotype) of STAP stem cells. d, Methylation status of the Oct4 and Nanog promoters. e, Electron microscope analysis of STAP stem cells. Scale bar, 1 mm. f, g, Beating cardiac muscle (mesoderm; 38%, n 5 8). Red line indicates an analysed region for kymograph (g). h, Clonability

of STAP stem cells. Clonal expansion from single STAP stem cells was performed. Pluripotency of clonal cell lines was confirmed by teratoma formation assay, showing the formation of neuroectoderm (left), muscle tissue (middle) and bronchial-like epithelium (right). Scale bar, 100 mm. i, Production of chimaeric mice from STAP stem-cell lines using diploid embryos. *These STAP stem-cell lines were generated from independent STAP cell clusters.

j, Production of mouse chimaeras from STAP stem-cell lines by the tetraploid complementation method. *These STAP stem-cell lines were generated from independent STAP cell clusters. k, No H3K27me3-dense foci are seen in female STAP stem cells (n 5 50; the CD451 cell is a positive control). Scale bar, 10 mm.

RESEARCH

ARTICLE

Extended Data Figure 9 | Effects of various stressors on STAP conversion. a, Percentages of Oct4-GFP-expressing cells 7 days after stress treatment. Somatic cells were isolated from various tissues and exposed to different stressors. Oct4-GFP expression was analysed by FACS. b, Oct4 and Oct4-GFP

expression induced in the reflux oesophagitis mouse model as an in vivo acid exposure model (top, experimental procedure). Oct4, but not Nanog, expression was observed in the oesophageal epithelium exposed to gastric acid (75% of 12 operated mice).

0 comments :

Post a Comment

Popular Posts

About

Blog Archive

Next

Newer Post

Previous

Older Post

0 comments :

Post a Comment