Reactions to RIKEN’s STAP cell protocol

A small group of the STAP authors via RIKEN have now released a detailed protocol also available at Nature Protocol Exchange for making STAP stem cells entitled “Essential technical tips for STAP cell conversion culture from somatic cells”. This comes about 5 weeks after publication of their Nature STAP stem cell papers.

2020 update: With the 4th and perhaps final Obokata paper retraction this year, is the STAP saga finally over?

Drs. Obokata, Sasai, and Niwa have added more detail to this protocol compared to the methods in the Nature article. They say further that a protocol paper will be published in a journal with even more detail later. A key question that popped to mind is why Dr. Charles Vacanti was not an author on this methods piece? Will he be an author on the actual methods paper? If not, why not?

The authors start off the protocol with some background summary about STAP and then a cautionary note:

“Despite its seeming simplicity, this  procedure requires special care in cell handling and culture conditions, as well as in the choice of the starting cell population.”

To paraphrase, I’d say the gist is, “Caution: STAP cells are extremely difficult to make.”

It would seem from this protocol (see quote below) that STAP is unlikely to work in a broad number of cell types even though the original paper sure seemed to highlight that it did work in many different cell types.

“The types of cells used for STAP conversion are also critical, and the use of cells from other sources (e.g., the use of cultured fibroblasts after passaging) may also result a failure to achieve STAP conversion. We have reproducibly observed STAP cell conversion when proper procedures are followed in  the correct sequence”

However, they do say that primary MEFs can be made into STAP cells.

The authors also reinforce their argument of a key distinction between STAP cells and STAP stem cells, which I think many did not fully appreciate from the original papers. So everyone has to be more careful with nomenclature here. STAP cells are not the same as STAP stem cells.

The first 8 steps in the new protocol create the “STAP cells”.

Then the authors list a separate procedure for making “STAP stem cells” or FGF4-induced stem cells “FI stem cells” after that from the same starting “STAP cells”:

“STAP stem cells, which are converted from STAP cells in ACTH-containing medium (see Procedure), lose the ability to contribute to extra-embryonic tissues. FI stem cells, which are generated from STAP cells in FGF4-containing medium, in contrast retain the capacity to contribute to both embryonic and extra-embryonic lineages in blastocyst injection assay, although their embryonic contribution is relatively low.”

So just to be clear they report that the first 8 protocol steps alone are reported to create totipotent cells.

Then to make STAP stem cells they list just 2 additional needed steps of growing cells in ACTH-containing media on feeders and then switch to ESC media with 20% FBS. It seems quite simple.

STAP cell generation from T cells?
STAP cell generation from T cells?

The oddest thing about the new detailed STAP protocol seems to relate to the T cell receptor.

Strangely, the authors seem to indicate in an “IMPORTANT” note after these 2 STAP stem cell inducing steps that in fact it seems that their STAP stem cells (and presumably STAP cells?) are not actually arising from T lymphocytes:

“We have established multiple STAP stem cell lines from STAP cells derived from CD45+ haematopoietic cells. Of eight clones examined, none contained the rearranged TCR allele, suggesting the possibility of negative cell-type-dependent bias (including maturation of the cell of origin) for STAP cells to give rise to STAP stem cells in the conversion process. This may be relevant to the fact that STAP cell conversion was less efficient when non-neonatal cells were used as somatic cells of origin in the current protocol.”

Doesn’t this contradict Fig. 1i (see above) of the original Nature article that showed T cell receptor rearrangement? Also the authors even in this new protocol paper refer to the cells of origin as “lymphocytes”.

I’m not an immunologist and perhaps I’m missing something so I’d be curious if others with a deeper knowledge of T cells could weigh in here on the meaning of the author’s statement vs. the original paper.

On a simple level to me this new statement seems like a red flag, but perhaps others can clarify the meaning and whether this suggests the authors are making STAP cells actually not from lymphocytes but rather from some kind of hematopoietic progenitor/stem cell.

OK, so what about the first 8 steps to make the STAP cells themselves?

Nothing particularly unusual struck me about it other than perhaps overall the numerous apparently crucial subtleties to the protocol.

I need to go through the new protocol in more depth and compare it to the original STAP paper, but to me this methods piece–even assuming it helps people independently replicate STAP and indeed it could well do that–would seem to narrow the likely impact of STAP cells/ STAP stem cells.

Jeanne Loring nicely summed things up in a comment on this blog a few minutes ago:

Now that RIKEN has released a detailed protocol, I hope someone can follow it and reproduce the results. I’m disappointed that it doesn’t work in mice older than one week and that the isolation and treatment of the cells has to be so specific. Since I’m not interested in reprogramming baby mouse cells, this is of no use to me. Does that mean that I believe in STAP cells? I may after someone else is successful after following the exact protocol. But do I care? No.

Did you all find anything else notable about this new detailed protocol? Other broader reactions?

78 thoughts on “Reactions to RIKEN’s STAP cell protocol”

  1. Dr. Obukata’s repeated silence to obvious questions about her work
    calls into question her total competence of what she is dealing with
    here – and maybe where it originated from. Dr. Vacanti’s statement
    that he will publish “his version of the protocol” is also curious.
    Together with the fact that I recently heard within the field that all
    of the idea and ground work behind this might have actually originated
    from the brains and hands of a third person present in Dr. Vacanti’s
    lab earlier, who left this work behind and from whom Dr. Obukata took
    over, things might start to add up. Apart from unsuccessful attempts
    all over the world to try to reproduce the claims in the publication,
    has anybody actually tried to trace the origins of this work – coming
    out of the lab of an anaesthesiologist and from a then PhD student?

  2. An immunologist Prof. Yoshimura of keio univ in tokyo , comments on absence of rearranged tcr allele at length in his blog, which i found is interesting.

    Also, the bio informatician mentioned above updates his blog today. He adds rearrangement of tcr was not observed cells after acid basing. So neither stap cells or stap stem cells are originated from T cells if he is right. He says some more interesting things.

    1. Do you mean the difference in mouse strains used for
      the Chip-seq input data?

      NCBI site data indicates similar things.
      As shown below, the cells other than CD45+ cells were derived from cag-gfp mice not from Oct4-GFP mice(*).

      CD45+cell(derived from Oct3/4::gfp C57BL/6)

      ESCell(derived from C57BL/6 x 129/sv )

      STAP Cell(derived from C57BL/6 x 129/sv)

      STAP Stem Cell (derived from C57BL/6 x 129/sv)

      (*)The protocol released yesterday says, “C57BL/6 carrying Oct4-gfp (29 of 29), 129/Sv carrying
      Rosa26-gfp (2 of 2), and 129/Sv × C57BL/6 carrying cag-gfp (12 of 16). STAP stem cells with all these genetic backgrounds showed chimaera-forming activity”.

      I’m not sure why they picked up these combination.

      1. Good point. Nevertheless, it argues even more for the hypothesis that their STAP and STAP stem cells are nothing but contaminating ES cells. Their choice of samples is illogical at best in terms of proving that STAP and STAP stem cells were derived from stimulated CD45 positive cells.

        To add another minor problem, they seem to have forgotten removimg RNA-seq and ChIP-seq protocol from the methods section in the Article paper.

        1. I agree with W.S. Their strategy is illogical and not clever enough.

          Right now, I’m just curious about the validity of the Chip-seq data described in nature paper.

          I’m not a specialist of this field and I might no be true. so if some of you know the technology very well, please let me know the answer.

          The cells other than CD45+ cells (derived from C57BL6 congenic Oct4-gfp) used the analysis are derived from C56BL6 X 129sv.

          I’m just curious about these two things as follows

          1) Is it possible to compare Chip-seq data between the samples from different strains?

          2) Is it possible to compare Chip-seq data between the samples from mixed background mice like C56BL6X129sv?

          As you know, there are multiple regions which exist in genome of one strain but not in the genome of another.

          I also noticed that the contributions of original strains in the genome of mixed background mice are different from mouse to mouse, especially when the mice are not well backcrossed, such as F2 or F3 mice.

          As you know, Chip-seq described in their paper is an assay to sequence the genomic DNA precipitated with anti-body against H3K4 and H3K27. I suppose that the minimal assumption of these kinds of assay is to make sure that the input DNA (same mouse origin, at least same genetic background) should be same to each other.

          Therefore, I think that it might be difficult to compare ChiP-seq data from multiple strains or a mixed background stain.

          I’m not sure if the region used in the papers (the region around Oct4, Nanog, Sox2, Gata2, brachyury, Nkx6-2, Cdx2, Eomes, Itga7) include the regions which are affected by the differences in genome between C57BL6 and 129 strains. But if so, their analysis of Chip-seq data is meaningless (If not, it might not cause a big problem).

  3. I am not a specialist in this field, but one thing that I know is that the primordial germ cells in some birds preferentially grow under an acid condition rather that at pH7.2 or so. One more would be that addition of some acids like butyric acid facilitates expression of genes exogenously integrated into chromosomes.

  4. Below is translation of an anonymous blog written in Japanese, seemingly a whistle-blower in RIKEN institute.

    Non-existence of STAP cell : diary of kaho
    15:10 March 5, 2014 diary by kaho

    This sucks. Riken’s response.

    They announced ‘Essential technical tips for STAP cell’, but it says nonsense. Because no STAP cells exist. I don’t say wrong ways of writing paper but literally ‘non-existence’. I’ve provided evidences but unaccepted.

    Many problems have been pointed out, including fabrication of figures and plagiarism, wrong interpretation of data. They would by themselves deserve retraction. But I don’t say them. They have been already discussed elsewhere, and I want to point more essential matter, i.e. non-existence of STAP cells.

    Why can you say STAP cells do not exist? I’m not an insider of the paper. So, I don’t know who made the mistakes and what s/he intended. But I can prove their fabrication, or complete falsehood at least. They spoke out their fabrication to the world unawares.

    From which data? See the raw data of the next-generation sequencing. One of the two papers performed ChiP-seq experiment. And the authors released the data some time after publication (it should have been done immediately, though). Briefly, this experiment reads chromosomal sequence as a control, named ‘input’. Since the data record almost random DNA fragments of the cell, you can reconstruct its chromosomal structure by looking carefully.

    The paper claims that initialization occurred by stimulating T cells with acid. The initialization was distinguished by rearrangement of TCR, from selection of pre-existing stem cells like Muse cells. TCR means T cell receptor, which is cut and ligated so as single cell to produce single kind of antibody. That results in difference of DNA length between T cells and the other cells, and you can see whether the cell has once experienced differentiation to T cell. Strangely, occurrence of the rearrangement in mice made from STAP stem cells was not described in the paper. People requested this data from the beginning of the buzz, but it hadn’t been presented.

    To this proper opinion, I’d been thinking ‘you can do it’. This is the ‘input’. Since this data is just around 50 basepairs, the whole sequence cannot be reconstructed. But, the rearrangement can be detected by loss of the cut-off sequence. Since genomic rearrangement shortens specific position of chromosome, a part of the sequence will be lost. If STAP cells are made from T cells, CD45+ cells and STAP cells are expected to have shortened TCR region compared to ES cells.

    When I started this analysis, I just wanted a light sense of superiority by finding something that cell biologists cannot do. The result was surprising: first, CD45+ cells exhibit slight TCR rearrangement, while STAP cells and low-pH-exposed CD45+ cells did not had the rearrangement. I thought my way of analysis might be wrong, so analyzed data from defferent T cells, and confirmed TCR rearrangement. In other words, STAP cells do not derive from T cells.

    At this point, I suspected contamination of stem cells by incomplete purification T cells. It was too optimistic. Low-PH treatment almost destroy T cells. Then, why STAP stem cells were claimed to have TCR rearrangement? It does not mean low skill of experiment or mistakes.

    The truth will be unveiled by comparison of the ‘input’. I’ll write on this matter another day because so long and complicated.

    Non-existence of STAP cell #2 : Diary of kaho
    4:27 March 6, 2014 diary by kano

    I’m surprised with huge responses to my last blog, beyond my expectation. As several people guessed, I would be ‘an insider’ from outsider’s view of the institute, although I am not directly involved in the papers of matter.

    Inside the institute, I act with real name, and there’s no intention of hiding myself. I want to solve the problem within the institute as much as possible, otherwise I would provide all the information to outside.

    The aim of my action is rapid withdrawal of the papers and pursuit of the truth as much as possible, with an anger to the damage to trust on science and the institute, and also a fear for possible restriction on my research activity that may be caused by leaving the matter obscure.

    Confident I am that scientific facts on my side, but not confident at all with political battle. I had thought of writing more evidences, but I will put off because the situation demands me to shoot second and third arrows without showing my strategy to the authors of the papers.

    The comparison data of ‘input’ described in the previous blog, can be viewed from the addresses below. As the blog cannot post a picture, use UCSC Genome Browser instead. Anyone can reproduce the same result using publicly open data.



    Here is a giveaway. What it means will be unveiled soon.


    1. Because this is an anonymous person and until someone who identifies themselves openly has done a genomics analysis of the STAP data in the public domain, the opinions and conclusions stated in the comment should be considered in that context and not overinterpretted.

      1. This is the first time in my life I wished I was more interested in hematology and immunology during the 8 years (non-sequential) that I worked in both fields.

    2. This attack on TCR rearrangement issue is way too over-interpreted. Even if this Chip-Seq data doesn’t support the rearrangement, it does not disprove the STAP cell. Remember, the paper claimed that they created STAP from liver, brain, muscle, fibroblast, pretty much everything. It will just prove that the origin of STAP made from CD45+ cells were not “matured T-Cell”. That’s it. You shouldn’t over-interpret the result.

      One should be critical on the sloppiness of the Obokata paper, but the sin of the original paper should not excuse this fallacious or exaggerated counter-argument.

      1. You have a point on over-interpretation. But on the issue of the authors claiming that pretty much any cell will work to make STAP, the reality is that the T cell claim was at the heart of their paper. Even in their new protocol from yesterday they are still flat out clearly saying that the protocol makes (and made for them in their paper) STAP cells from lymphocytes.

        1. Both original paper and the new detailed Obakata protocol is filled with ignorance, which is causing problem in our discussion. Majority of CD45+ cells from 1 week old mouse are granulocyte mainly neutrophil, also monocyte, so they shouldn’t have called it “lymphocyte”. Nature Reviewer should have corrected this before publication. Even among lymphocyte (T, B, NK cells, both immatured and matured form), matured T cell is not majority at 1 week mouse. As STAP cells are mixture of these cells, and PCR amplify the TCR, while Chip-Seq doesn’t, it’s difficult to conclude from Chip-seq data.

          In any case, if the mouse result is genuine, it must be true that either STAP reprogrammin is real, OR 1 week old mouse contained totipotent cell in many places of body, which would be amazing too. If the mouse result is not true, this would be one of the biggest fraud/hoax, as it involves *deliberate* contamination of “STAP cell” with ES/iPS cell.

          1. I agree until you use the word “deliberate.” Please expound on your reasoning in deciding to use the word “deliberate.” I am not being contrary, I wish only to understand the logic, as I agree with almost everything previous to this word. At what point do I cease to understand the scientific justification for the use of this word?

      2. redsoxfan people… BAWSTUN… HAHHHHH-VAHD.

        eyes open, my fellow scientists. there are many-a-snake running around. i expected harvard to send some of “theirs” to question the challenges being posed to this research.

        be critical of what you read!
        the proof is above.

        1. Not being from Harvard, I am none-the-less appalled at the nasty and ultimately libellous attacks on another Institution.

          Can someone tell me if these opinions have any basis in reality?

          1. A and B conversation, C your way out.

            the above is an illustration of high school mentality. if it doesn’t involve you, stay on the sidelines and form your own opinion of me (my handle doesn’t make me as anonymous as you’d think)

            you’re what they call a “busy body”. i am one too, but at least i’ve got something to substantiate my seemingly irrational and uncooperative behaviour. you’re not allowed to be a “busy body” until you do something important. i have nothing better to do than clean up the area of science.

            clearly you’re appalled and upset at the effectiveness of my ineffective and angry-worded rants. i am sorry you feel that way. i am sorry for the people reading this.

            i will abstain from further comments on this thread to optimize others’ viewing experience.

            ’till the next thread.

            1. I understand neither your response nor at whom it is aimed. I was not attacking you or anyone else. Just trying to find out if all this bashing or Harvard and Riken as quasi-criminals is based in reality.

    3. A follow-up has been posted.

      Non-existence of STAP cell #3 : Diary of kaho
      2:41 March 7, 2014 diary by kaho

      Unfortunately, I don’t seem to be able to get political victory.

      Here I want to emphasize to my readers that this is not lonely battle, however. All the people whom I talked with directly agree and support me, so my side overwhelm in number. In scientific journals, retraction from author side needs approvals of all the authors. It is the current system that there is no way to force withdrawal when an even single person persists.

      Well, their newest story is that STAP stem cells lack rearrangement of TCR while STAP cells have. As you can see from the TCR region in my previous blog, genomic reconstruction almost disappear when the cells are stimulated with acid. Most of the cells should be rearrangement-free, although a small population may have. In this regard, the authors should explain if they will make a corrigendum. They should survey the rearrangement of not only STAP stem cells but also STAP cells selected by Oct4-GFP expression. To match with the NGS data, the expected pattern would probably be different from that in the paper.

      Next, the explanation of the giveaway. It doesn’t mean the greatest discovery in history of ‘sex change by acid stimulation’, but a series of experiments has diverse genetic backgrounds. You can see clearly by zooming in to Oct4-surrounding region.

      In other words, only the CD45+ cells derive from transgenic mouse carrying Oct4-GFP, and the other cells from different line. GFP sequence was found in the latter, therefore they are supposed to be CAG-GFP cells which appear in the paper, constitutively fluorescent.

      It is quite a poor experiment design that adopting diverse origin of the cells in order to demonstrate difference between the acid-stimulated cells from the original CD45+ cells. It would mix up possible cause of the difference, acid treatment or genetic diversity. I guess that Prof Yoshimura in Keio Univ, who cited my opinion in his blog and I’m so thankful, would scold “It’s a basic for researchers to take firm positive and negative controls”.

      It is hard to understand why the authors used cells of different sex and DNA. Since experiments consume much money and time, there must be certain reason(s) to do so. For example, if the success rate of STAP stem cells establishment is very low, they would have to analyze the rare specimens.

      In the paper, however, they collected CD45+ cells, applied stresses, and exposed to special environment. The most abundant is the original CD45+ cells. What they did is that, substituting this cheapest cells to different ones, and claiming that the original CD45+ and the stimulated cells are different. Should be different, since the cells of different sex and DNA are analyzed. So what?

      There is a guess in my mind, but unfortunately I could not get a definitive evidence due to lack of time. My political defeat is not only brought by the consequence of power game, but also my capacity shortage.

      1. I’m not sure if the analysis of “kaho” is correct.
        But the information in the NCBI site tells that origin of the CD45+cells are different from those of the other cells (STAP, STAP-SC,ES) as below; the cells other than CD45+ cells were derived from cag-gfp mice not from Oct4-GFP mice(*).

        Quite interestingly, the third analysis of “kaho” is quite similar to that NCBI information.

        The other two analysis may be true, especially for second analysis. It is quite reasonable that the different sex chromosomes observed in CD45+ cells were observed since the origin of the cells (mouse strain) was different from those of the others.

        As for TCR reconstitution problems in Chip-seq data, I’m not sure if it is fair enough to compare CD45+ cells (B6 background) and ES cells (B6X129) as a positive control, though “Kaho” also showed the different positive control).

        Anyway, the differences in genetic background in Chip-seq data made the things more complicated.

        Here is mouse strains described in the NCBI site.

        CD45+cell(derived from Oct3/4::gfp C57BL/6)

        ESCell(derived from C57BL/6 x 129/sv )

        STAP Cell(derived from C57BL/6 x 129/sv)

        STAP Stem Cell (derived from C57BL/6 x 129/sv)

        (*)The protocol released yesterday says, “C57BL/6 carrying Oct4-gfp (29 of 29), 129/Sv carrying
        Rosa26-gfp (2 of 2), and 129/Sv × C57BL/6 carrying cag-gfp (12 of 16). STAP stem cells with all these genetic backgrounds showed chimaera-forming activity”.

  5. Mitsuhiro Yanagida

    The authors argue that TCR rearrangement occurred in STAP cells or sorted oct4-GFP cells in fig 1i, but such cells were lost from STAP stem cells during conversion from STAP cells into STAP stem cells.
    I think this argument does not contradict with first their statement in the Nature paper.
    Cells with TCR rearrangement may show slow growth rate, and those cells may be lost during proliferation due to slower growth compared to cells without TCR rearrangement.

    1. Dr. Yanagida,

      I always respect your attitude to science.

      However, regarding that point you raised, I disagree to what you said.

      What is a STAP cell? Based on what is written or what you claimed, they are acid -treated splenocytes that express Oct4, not necessarily pluripotent. It may be true that T cells survive the harsh acid treatment and acquire Oct4 expression, but it does not mean that T cells become pluripotent. Oct4 as well as Nanog is a good marker for stem cells, but this does not mean that STAP cells are stem cells. This is a good example of “true-true-unrelated”.

      I think Obokata is to blame, but there is much more fault with Editors in Nature, who intended to publish exciting, but not scientifically rigid stuff.

      1. If the authors insist ”we have established multiple STAP stem cell lines from STAP cells derived from CD45+ haematopoietic cells”, evidence of the presence of STAP cell lines should be shown. You know, RIKEN recently published a paper entitled, “A simple and highly effective method for slow-freezing human pluripotent stem cells using dimethyl sulfoxide, hydroxyethyl starch and ethylene glycol” in PLOS ONE. Even if Obokata did not use the new method, STAP stem cell lines must be preserved.

      2. MItsuhiro Yamagida

        Dear T.E.,

        Do you understand the difference between STAP cell and STAP stem cell?
        You should read carefully two papers in Nature.

        As you say STAP cell is not a stem cell, since STAP cell cannot proliferate in a medium for STAP generation.

        STAP stem cell is generated by a ACTH-containing medium.

        1. Dear Dr. Yanagida,

          STAP stands for stimulus-triggered acquisition of Pluripotency or something similar. To prove pluripotency, the authors need to show that STAP T cells can become STAP stem cells or that STAP T cells contributes to different germ layers of teratoma. Given that the former possibility is dead now maybe due to slow proliferation or other reasons, the authors need to prove the latter, though it was not demonstrated with the data presented. They might be able to show that a multi-gene signature “suggesting” pluripotency is induced as shown in figure in the paper, it is not a proof unless they run a single cell PCR to examine co-existence of such signature and TCR rearrangement or start the experiments with purified T cells. If they try the latter, T cell-derived STAP cells would be outcompeted as you suggested, or they would not have to insert a lane without GL band.

    2. You’re not real Dr yanagida but a pretender. If you are truly Dr yanagida, you should know what sty1 gene is. Dr yanagida is rather a yeast person and doesn’t know much about stem cells.

      1. I believe he is Dr. Yanagida unless you are Dr. Yanagida and criticizing the pretender. But it does not matter for discussion here.

  6. According to the news media in Japan, Dr. Obokata recently succeeded to reproduce the results by following the exact protocol as published. It is confirmed by Riken CDB so it’s not a tabloid:

    I am rather surprised by the fact that she is allowed to run experiments when the paper is under investigation, but in any case, I encourage everyone who are trying reproducing the results to contact Riken CDB asking for material samples from this reproduction.

    1. I’m not surprised and it seems reasonable that she continues the STAP work at RIKEN. There is no definite evidence that I am aware of that she personally has done anything wrong as an investigator at RIKEN. I believe that she deserves to be presumed innocent of any alleged misconduct unless conclusively proven otherwise.

      1. I totally agree with Paul, and I am appalled by the 6 people who rated down his comment (and the 14 who likes Thomas comments). What has Dr Obokata done that made her “guilty”?
        Trusting the pictures sent by her co-workers were correctly labelled? Trusting the method written by her co-workers and sent the her? Maybe misguided gel picture editing? English mistakes (I dare any of you to write a whole manuscript in flawless Japanese or French!)? She might have made mistakes under pressure due to being unexperienced (she is 30 for god’s sake, extremely young for a PI by Japanese standards), but she will never, never deserve your scorn and hate. Of course it is normal she continues to work on STAP, how else could she show it is the real thing?
        What you would have her do? Shackle her or something?
        Really, sometimes it seems many bloggers here want to see her fail, maybe out of envy of having made a breakthrough they were not clever enough to make themselves?

        1. Multiple pictures in at least one of the STAP papers (the article) are unambigously manipulated. The most problematic pictures for me (which presumably are Dr. Obokata’s) are the mosaic images of Oct4+ clusters of cells, made to look as if there are more (or less) positive clusters. This panel in Sup Figure 2, for example, has been clearly, demonstrably faked. This hasn’t caught the press’s attention as much as the placenta switch, but it was what convinced me that this paper needs to be retracted.

          Dr Obokata should carry on working while the investigation is in progress, and certainly deserves no hate. However she may have wasted an awful lot of people’s time, which isn’t OK.

        2. John, I haven’t said she is guilty. I agree with you and Paul that, unless there is a concrete evidence, she should not be considered guilty.

          However, think about this: given the situation that both Nature and Riken are investigating the validity of the results, is reproduction of the results by herself very convincing? I was just surprised by the fact that the Riken allowed her to take such a bad approach to prove the results (perhaps my wording was unclear).

          I agree with you that she should probably be allowed to continue working on STAP, but not as a way to prove the results under investigation. For her own sake, it would be better if someone external can independently reproduce her results.

          1. Of course I agree with both of you there are issues with both papers and previous reports that undermine the credibility of the research.
            My comment pertained to the attitude displayed here. There has been no official declaration that these data were intentionally tempered with. Can you say, for sure, that these faulty pictures absolutely prove STAP cells are a fabrication? Without investigation and proper “trial”, Dr Obokata and/or the other authors of both papers, should be considered innocent.
            As for RIKEN, all they want at this stage is to protect their own employee and reputation, so of course they will give the message to the public that, by the way, she still can make something that they believe is a STAP cell as described in their paper. It does not weight a lot scientifically, but it gives the message that RIKEN protect its own people. Without definitive proof that STAP cells are a fabrication, again, it is an understandable position to take. Nobody would want to work for an institute that fires its researchers based solely on hearsays and social medias.

    2. I read this article by asking my Japanese co-worker. This article does not provide any new evidence at all. And worst of all, they asked Obokata herself to “reproduce” the result. Reproduction experiment should be done by someone except her, using independent location, independent reagents. This is nonsense.

  7. I was also surprised to see Dr Obokata suddenly being not corresponding author despite the aggressive “she alone created the method” from the previous media coverages.

    As for Dr Vacanti’s vacant name, I guess he went looking for greener Pasteurs….

    Ok, more seriously. I noted that the acid treatment by HBSS start at 4C, and cells are then placed in the incubator. This was not mentionned in the two previous papers but seem rather important. I guess none of the scientist who tried to reproduce the method so far used HBSS at 4C.
    Instead of pH (+ trituration) stress we have temperature shock + pH + trituration stress = STAP.

  8. Well, being only an MSW, I was set kind of adrift in the sea of data here. 😉 But it has made me think of one point. I’ve thought for a while that we might need another choice on the STAP poll: “do you believe that STAP cells are anything close to what they were claimed to be?” Because it certainly seems like that’s a question which can be answered with a pretty unequivocal “no.”

  9. Given some details as useful tips, I got several things sounding odd…

    There are several steps potentially causal for contamination of non-targeted (undesirable) cells. However, for example, there is no information about the gate setting of FACS sorting. I do not think the purity of CD45+ cell would be 100%, which depends on the setting of FACS sorting and also requires data of re-analysis for the purity. And, one thing that surprised me is that the STAP cell line is not derived from monoclonal population but from pooled cells of multiple colonies…..Imagine what this means….

    Furthermore, in the protocol, they report that they have performed 50 times of the same procedure, which is a lot. Amongst these 50 trials, they succeeded to obtain STAP colonies for 40 times. Then, I do not think that they could perform all these trials with the consumables that they claim that lot-to-lot variation would matter…. Personally, repeating that sequence of experiments for 50 times in a few years would be impossible…. But, this is not the principal point….

    1. Repeating the experiment for 50 times is not unusual at all. You shouldn’t use the word “impossible”. Anyone in iPS lab can easily do 50 reprogramming experiment in a few years, especially if they can outsource the making virus part and passing cell line part. Time consuming part is the passage of cells, but in the case of STAP, it seems they don’t spend much time on creating long passage cell line.

      1. Note that her lab is small unit with limited resources.
        And, considering the procedure, they claim that the cells should be prepared freshly. Then, it should have required mating of mice to obtain new born pups, FACS sorting appointment, and so on. Practically, I think that the whole procedure would take 3 weeks including all side works such as re-analysis of sorting quality and so on… Then back to the comment of Dr. Wakayama, the concrete protocol should have been established in these 2 years or so. Besides all trials with other cell types, repeating 50 times of that full procedure would be incredibly high burden in terms of both man power and budget. Also note that, if someone other than Dr. Boo kata had performed certain portion of those 50 trials, reproducibility should be proved by the testimony of this person. But, nobody has raised a hand for this, implying that those trials should have done mostly by Dr. Obokata…

        1. In addition, given all the “not to do” cautions in the protocol exchange, tremendous amount of try-and-error efforts have been made. All these should be based on every single full procedure that failed in the end due to one “not to do” tip. Presenting the tips with clear “no!” should indicate a lot of experiments which take time and cost substantially. All these efforts are now reported to be done by Dr. Obokata who should be able to use only 24hours for a day equally as us…..

          1. Remember that they have been working on this for years and also more recently Dr. Obokata has had her own, likely amply staffed lab and wonderful shared core resources at RIKEN. It’s doable I think to have done all this work.

      2. I agree. It’s not impossible at all. It’s also worth noting that multiple staggered STAP experiments can be done at the same time in an overlapping manner too. For example, Obokata’s lab could have had a half-dozen STAP experiments going just during a week or two.

        1. I am Japanese scientist and know well about the inside of RIKEN. I do not think that Dr. Obokata’s lab got so much extra funding for personnel because of that project.

          Again, considering the current situation, if someone has done the procedure independently, RUKEN should have announced the fact in the way stressing it.

          Also, managing that procedure in overlapping way by a small number of persons besides other works is very difficult.

          To obtain new born pups in a timely manner is also not easy as the appointment of FACS Aria which should be always crowded and tight busy should be timely arranged. I am thinking it very practically and would say that it is not easy as imagined to plan that kind of experiment in an institute where core facilities are shared by number of labs.

    2. Coming back to scientific discussion, CD45 labels multiple types of cells of mainly hematopoietic lineage. However, those CD45+ cells differ in their size. So, if they used only CD45 to discriminate T lymphocytes, they should have set gating for a specific size of population. This is very important as they claim that they separate homogenous population and should be recognized well by them as an important condition…

      Another thing, in the protocol exchange, there is one step where cells are suspended in H2O. Considering the small size of lymphocyte, the pellet of one million T cells should be very small in volume. Suspending it in a relatively high volume of H2O, even for a short moment, would kill the cells for sure… But, surprisingly, they recover the cells without any trouble…

      Also, it is strange that they get red blood cells in the suspension even after sorting….

  10. Mikio Hirayama

    I received the response today to my inquiry of the trituration procedures from Dr Niwa who is the correspondent of Protocol Exchange. The answer was as follows. As he
    says the technical tip is just minimum information to support reproduction, it seems to
    mean that it is still very difficult to reproduce their experiment!

    Answer from Dr Niwa:
    We will publish the full protocol in near future, which will cover all informations.
    The present version of technical tip is just minimum information to support reproduction.

    Best regards,


    On 2014/03/05, at 20:33,
    > Dear Hitoshi Niwa, M.D., Ph.D. (
    > )
    > I appreciate very much if you answer my question below. I sent the
    > following e-mail to Dr Obokata. In your published today’s protocol, you
    > did not describe the trituration procedures in detail. Is it not important?
    ( I omit the details of my question, because I posted the question in this blog.)

    From Protocol Exchange
    Essential technical tips for STAP cell conversion culture from somatic cells
    Haruko Obokata, Yoshiki Sasai and Hitoshi Niwa
    “To facilitate the broad testing and use of this technique, we are now preparing a full protocol article with step-by-step instructions. However, as the preparation, submission and publication of a full manuscript takes a significant amount of time, we would like to share a number of technical tips for STAP cell conversion culture (and related experiments) in this Protocol Exchange. We hope that these technical tips may answer many questions frequently asked about the experimental details.”

  11. I’m increasingly convinced that their technique is just a way to select stem cells from tissues at development. The restricted use of one week old mice turn its very clear. By adding acid, they are just killing normal cells and stem cells from the niche somehow survive, possibly stop to differentiate and show signs of pluripotency.
    I’m almost sure that the muscle cells they report to reprogram are just satellite cells from muscle niche, as at this age they represent 35% of muscle fiber cells and are pluripotent.

    1. In both the original and the new and improved protocols, I do not read that the protocol is restricted to neonatal cells. What they do say is the EFFICIENCY decreases with older cells.

  12. I read very interesting analysis by bioinformatics professional in RIKEN.
    Analyzing the input of ChIP-seq data published, both STAP and STAP stem cells had no TCR-alpha and beta rearrangements.
    More surprisingly, original T cells for STAP cell production were female, but STAP and STAP stem cells were male.

    TCR alpha

    TCR beta


    1. I am not sure if this is true. The original paper used both male and female mouse for STAP cell, sometime they used mixture. I couldn’t be sure for 100% that the Chip-Seq data came from female. It could be male. The method section doesn’t say that clearly. (In general, this paper is a mess.) If you have this information (sex for chip-seq data), could you post it? Otherwise we shouldn’t jump to conclusion.

    2. I re-checked the paper, and it says:
      “Because the number of CD45 cells from a neonatal spleen was small, we mixed spleen cells from male and female mice for STAP cell conversion.”. So perhaps they used both male and female. Also they never isolated T cell and reprogrammed “T cell”. They reprogrammed CD45 cells which is known to contain hematopoietic stem cell and other progenitors, and they conjectured that it must have contained some T cells based on the PCR data, which is probably due to contamination.

      What’s odd to me is the part they claimed they easily created STAP from all other tissue relatively easily. This completely contradicts to what they now claim (STAP is extremely difficult to make.) .

  13. You’re very right their new information contradicts figure 1i. It more directly contradicts the text of the main STAP paper, which says:

    “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. “

    1. Perhaps I am missing something. Somebody please help me understand what is going on: they derived stem cells from the T-cells as indicated by TCR rearrangement, after which immune response to new antigen becomes impossible. And they show fully healthy embryonic chimera without immunedeficiency?

      1. T cells are not necessary for embryos or adult in SPF. There are many T cell deficient mice which are viable and breedable, and Rag-deficient mice, lacking both B and T cells, are healthy without infection.

      2. No, that’s not quite the issue – TCR rearrangement is part of Tcell development – all other cells in the body have a complete

        But upon maturation T-cells recombine that region through a process known as VDJ recombination, which results in a irreversible loss of some DNA around that region, which is why the bands in lane 3 in figure 1i are shorter. STAP cells derived from recombined T-cells should therefore only have the shorter, recombined bands, not the intact band seen in the germ line (control lanes 1 and 2). Instead, the STAP cells in figure 1i have both recombined and unrecombined bands (lanes 4 and 5), which makes the image internally inconsistent.

        Regardless, in the text of the paper the authors state that they observed the recombined bands in their STAP cells, “indicating at least some contribution from lineage-committed T cells.”

        According to since-deleted reddit posts from a someone claiming to be a member of Charles Vacanti’s lab, the authors made STAP cells from the T-cells because the original reviewers wanted bona fide proof that they were indeed reprogramming cells, rather than isolating a small heretofore unknown population of stem cells present in the original tissues. Working with cells that did not have a complete genome, and finding that same incomplete genome in the STAP cells, was therefore conclusive evidence that they were reprogramming cells.

        So, to see the authors now claim that none of their STAP cells carry the recombined version of TCR is rather frustrating, to say the least!

        1. It’s important to understand that the stap cells supposedly only exist in clumps, that likely include STAP cells derived from both Tcells and non-rearranged cells. STAP cells are called pluripotent because they express Oct4 and other markers and can make a mouse via tetraploid complementation (the video of a green embryo).

          STAP stem cells are a different animal, generated after culture in ACTH containing media. We know know that STAP cells from rearranged VDJ-Tcells, do not form STAP stem cells, or only at a undetected low efficiency.

          technically, this could mean that Tcells form oct4 expressing STAP cells that are qualitatively inferior to other cells in the STAP cluster and will not reprogram into STAP stem cells.

          Importantly, we need to know wether the other test for pluripotency, forming that green embryo was passed by T cell-STAP; does that green embryo have VDJ rearrangement in all its cells? (and thus monoclonal lymphocytes)

          I suspect it doesn’t since that would have been such an obvious thing to show. Thus, realistically, this probably means that they did not reprogram T cells to a functionally pluripotent state.

  14. CD45 is expressed on lymphoid, myeloid and some hematopoietic stem cell. They should have used more committed marker such as CD3. In addition, FACS sorting is not perfect process. It’s difficult to have 100% single cell suspension and you often end up with some contamination.

    The most straightforward interpretation of Obakata paper is that, CD45 sorted cell population contained some totipotent stem cell. And this population could be very rare and may be difficult to reproduce it. And T-cell rearrangement was due to contamination of DNA from committed T cell in the original mouse. Reviewer should have required quantitative real time PCR result, instead of dirty gel picture. There is no reprogramming.

    1. Exactly- 100% agree. As an immunologist, the whole paper from the start has seemed like a rare-selection protocol rather than a reprogramming event. The acid-treatment selects for a population of CD45+ totipotent cell that might even have been inadvertently sorted for due to autofluorescence. The lack of TCR rearrangement admitted to in the technical notes and the use of a lousy marker- CD45 all point to a selection event. If that’s the case the discussion / conclusions in this paper are erroneous. While I don’t think the paper is necessarily fundamentally wrong- I do not think it will live up to the hype.

  15. Quote from the protocol:
    “FACS sorting can be an important step for the confirmation of cell purity, but can affect both cell viability and reprogramming efficiency. Skipping this step may increase reprograming efficiency ,although this may result in a reduction in confidence in cell identity.”

    I’m not a stem cell expert at all, but this does seem to present an almost Schroedinger’s-cat-like conundrum:
    If I don’t know for sure which cells I’m working with, I can make STAP cells. But if I look inside the box, the protocol doesn’t work (or only rarely). The statement also may suggest that the authors routinely skipped this verification step.

    Numbers would be intersting here: How often did the protocol work with FACS-verified cells, how often did it fail?

    Please correct me if I’m wrong… just my lay-observer’s impression.

  16. @Jeanne Loring
    I sit at the same table in that I am not interested in reprogramming neonatal cells, but rather stinky old rotten ones 🙂

  17. Regarding the statement on T cells not giving rise to STAP- I think it doesn’t necessarily contradict the figure 1 gel. We’ve discussed how the STAP clusters are polyclonal, and include non-rearranged cells giving rise to that upper band. It could be that those non-rearranged cells (B cells or whatever didn’t sort out properly) make STAP that are more amenable to convert to STAPsc’s when put in ACTH or FI media. Sure- not impossible. However, this was one of the key points I was looking forward to; demonstration of rearrangement in the STAP scs and the cloned mice that would finally link the mature cells to the functional and indisputable assays. Now we’re further away from excluding ESC contamination!

    1. Now, STAP cells and STAP stem cells are different. Then, how cold they say that STAP cells acquire pluriotency?

      If I believe what they have shown (by ignoring data fraud, if any)
      1) new born a pool of splenocytes, including T cells, can turn on Oct4-GFP or Oct4 mRNA after acid bathing.
      2) some cells had possessed or acquire pluripotency.

      Aren’t they “true-true-unrelated”?

  18. It is getting more and more sketchy. Fig.1i in the original paper only means that some T cells are there or some DNA from T cells was in their DNA prep, although I am not sure if they purify Oct4-GFP+ cells or not. If not, it is not surprising. But what is critical is that T cells do not contribute to STAP stem cells, indicating no evidence of reprogramming of mature T cells. And again, no GL band visible in the inserted lane (even despite PCR disadvantage for long product) looks so artificial if they started only with CD45+ cell purification or enrichment. 1 week-old mice would have 10-20% of T cells in the spleen.

    In the method release, they still keep saying that STAP stem cells contribute to both embryonic and extraembryonic tissues, but this must be proven by a single cell injection. If FGF treatment induces conversion of some ES cell type cells into TS-like cells, making up a mixture of two different types of “stem cells”, the same outcome would happen. I think this is an immature conclusion. Please correct if I am wrong.

    The important point for the two original papers were:
    1. Easier to make than iPS.
    2. T cells can be reprogrammed.
    3. Bipotnetial for embryo proper and extraembryonic tissues.

    How many of these are viable now?

    Why is Dr. Niwa a corresponding author for the method release??? Missing Vacanti is weird but this is also weird. Obokata should have been the corresponding author.

    1. I will avoid commenting on the legal and ethical concerns of the STAP phenomena and the inconstancies in what has been claimed and published to date…and I agree with most all of them.

      None-the-less, I am stupid enough to have keep trying the published method. The “New and Improved” version was not much different than what I ended up doing: I was divining what the protocol probably was, but not published.

      I had no young animals and suffered the wrath of those who found this pointless. We worked with the tools available to us.
      Not every lab is outfitted by royalty.

      Both protocols mention that it works in older cells, but with reduced efficiency. I am of the reasoned opinion that efficiencies of 3% to 0.001 %, during all sorts of dedifferentiation, are far more meaningful than the commercially driven desire to have everything be 100% efficient: 100% efficiency is not an apodictic biological reality.

      Anyway, We have tried the protocol 2x on Lympocytes from 6 month old rats, once on MEFS, twice on lymphocytes from an Oct4-GFP knock in mouse, and once with lymphocytes from my own veins.

      1x we were able to detect Oct4 in the treated rat lymph’s vs untreated.
      1x we were able to detect CD45-/GFP+ in treated vs untreated, by flow cytometric analysis .

      I am not saying it works, yet, just my experience using what tools we had.

        1. Hi, this looks encouraging indeed. Would you be able to upload the data on the STAP new data page, so that it is available for everyone to see? Or do you consider it is too preliminary to show…This could give a boost to people trying to replicate the work.
          In general every protocol includes many cooking tips, and one should always be very precise…this is a common experience in biology, and it may explain why different labs have failed up to now. Once the relevant parameters/ essential steps are identified, one can deviate from the protocol, but they are clearly not known at the moment.
          There may be fundamental reasons why older cells are refractory to reprogramming- if the story is real- but even being able to reprogram newborn cells could lead to fundamental clues for adult cell reprogramming, and maybe provide insight into other phenomena, such as cancer. Your result in older mice are encouraging.

          1. I am a little hesitant to post the data yet, but I have shared it with the blog administrator to aid in keeping my posts believable. When i have enough data, I´ll share it.

      1. With respect to T cells, please may I remind the readers to the normal in vivo state of T cell quiescence until activated through the TCR: i.e. T cells are in G0 of the cell cycle. I find it strange that the published STAP cell protocol for T cells apparently works without activating their entry into the cell cycle.

        A second point of note is that Obokata et al (Nature 505, 641–647 (2014)) implied that CD45+ cells never express pluripotency-related markers unless reprogrammed: this is misleading. Adult CD4+ T cells when activated have been reported to express Nanog: this occurs when the T cells lack MARCH7, an E3 ligase required for degradation of the LIF-R (Cell Cycle 9, 4213-4221: 2010). This might be helpful to those looking at the Oct4gfp mice.

    2. Correction: They say that “STAP cells are also found in extraembryonic tissues, such as placenta” and that “SATP stem cells… , lose the ability to contribute to extraembryonic tissues.”

      I agree they still must prove that STAP cells differentiate into trophoblast derivatives in vivo. Their experiments as they are have several problems. 1) Their control chimeras in figure 1: In chimeric embryos between E9-12.5 (the stages they showed) made by injecting ES cells into blastocysts there should always be some contribution of ES cells derived extraembryonic mesoderm to the placenta and fetal membranes, however none of their ES cells derived chimeras showed that feature. That is very hard to explain.
      2) Since they “didn’t find” GFP positive cells in E9.5-12.5 placentas of control chimeras they just assumed that STAP cells-derived GFP positive cells in the placenta are of trophoblastic nature when those cells could be GFP positive extraembryonic mesoderm. No colocalization with specific trophoblast markers was performed in extended data figure 1b. So the presence of GPF positive cells in that area doesn’t prove trophoblast differentiation of STAP cells. 3) Likewise, since they “didn’t find” GFP positive cells in E9.5-12.5 yolk sacs (fetal membranes) they just assumed that STAP cells differentiated into visceral endoderm cells and use a pan-cytokeratin antibody as a marker, even though is not an exclusive marker of visceral endoderm cells.

      The pictures in extended data figure 1b and c do not show the context. In which area of the placenta/embryo were those images taken and at which stage?. Could it be possible that the STAP-derived GFP positive cells shown are mesodermal cells? Or definitive endodermal cells? Maybe they just over-interpreted their chimera experiments results?

      In summary, the 3 important points for the two original papers that you mentioned are not sustainable in my opinion.

  19. If it only works with cells that haven’t been expanded in culture, that might go a long way toward explaining the negative replication attempts so far. I’d bet most of the people who tried with MEFs or human fibroblasts used culture-expanded cells.

    1. True. Also though it just seems overall that the types of cells for which STAP can probably work is getting narrower and narrower. I guess we should keep an open mind as this evolves.

    2. i think this is cheating and abuse of the postpublication process.

      RIKEN is clearly acting on behalf of Harvard, and is trying to now change the original claims in the article after-the-fact.

      i still feel a retraction is necessary, because these people are now trying to change what was PUBLISHED in order to KEEP their publication. that is not fair.

      furthermore, they’re now making all sorts of qualifications that WERE NOT in the original work or even SUGGESTED. how is this procedure even considered as acceptable? it doesn’t address the issues, which were to live up to the claims being made in the article.

      it’s like getting caught for cheating by your teacher after you wrote in pen, and then getting a re-grade after getting your test back and changing the answers.

      this is so pathetic and high school. but i expect no less from the cliquey individuals at harvard, they’re the kids I used to make fun of in high school IB class. they thought they’d be the ones who got the last laugh, as i sat there mocking them for taking high school so seriously. it’s no wonder how they ended up as academic slaves on the board of the most prestigious journal in the world.

      this is nonsense and unacceptable.

      1. Actually I think it is more likely that Harvard and RIKEN are opponents on this at this point just agreeing on one thing and that is that they hope that STAP is the real deal, but on what comes next from STAP, etc. they are not buddies.

        1. i agree with you dr knoepfler in that they are going to be opponents from hereon out.

          nevertheless, it took quite a bit of gall to publish something that presumably was going to need reproducibility to verify the “impact” and reach of the conclusions.

          at this point i’m having a hard time believing Harvard wasn’t the larger investor in all of this (given that the 2004-2005 papers are showing signs of forgery/misrepresentation, and shortly after this author relocated from their lab in Massachusetts to Harvard). nonetheless, we’ll see if the cards fall as i think they will with respect to who is orchestrating this. note that i don’t doubt the institution has done good things before, but as i’m sure you would agree dr knoepfler: the time of dining-out on insitutionalized academic credibility has past.

          i would think that the reason the institutions are now opponents is precisely due to the unexpected controversy and discussion. i firmly believe that whoever put Nature’s editors to publishing this, thought that the current controversy would subside and not require retraction.

          at this point, i’m concerned about anything OTHER than a retraction because the investors are likely going to wait as long as they can. with the publication secure, this controversy can be forgotten in a few short years, and the paper can be re-trotted out to gullible investors (for God knows what reasons).

          nonetheless i swore to not post (sorry!), but you did pose an interesting viewpoint that i agreed with (even though we likely diverge on many other points surrounding it!)

Comments are closed.