Historic turning point for IPS cell field in Japan?

As many of you know, the pioneering, first of its kind IPSC clinical study in Japan has been suspended as I first blogged about here. 

In the comments section of that blog post there has been a helpful overall discussion that has involved Dr. Masayo Takahashi, the leader of the trial. It is great that Dr. Takahashi has been participating in this discussion and I commend her for that openness.RPE sheet_pr

This comment stream has been particularly important because the media have only minimally reported on this important development. There have been only a few articles in Japanese (several months ago) and as far as I know only one in English, which was posted in the last day or so in The New Scientist. Unfortunately The New Scientist article, as many have noted here, used an inflammatory title invoking a supposed “cancer scare” and some over-the-top language. Although that article had some bits of important info, the negative bias in the article made it overall not very helpful. Some readers of that article were likely confused by how it was written and the title.

The clinical study in question is for macular degeneration and involves the use of sheets of retinal pigmented epithelial cells (RPE) made from IPSC (e.g. see image above from RIKEN). Several of us have been discussing the suspension of this trial over on Twitter too including Dr. Takahashi (@masayomasayo). Some tweets by the community have been constructive. Others not so much.

Two main possible issues have come up in the discussion of the reasons for the trial stopping: (1) six mutations were detected in the 2nd patient’s IPSC and (2) significant regulatory changes are on the way in Japan that apparently in some way will delimit IPSC research there. Dr. Takahashi has indicated that the latter reason was the dominant factor in their decision to suspend the trial. The fact that the 2nd patient’s IPSC reportedly had six mutations that were not present in the original somatic cells warrants further discussion too. For example, when and how did these mutations arise? To be clear, however, I do not see (based on the information available) that there was a “cancer scare” by any stretch of the imagination as The New Scientist article had indicated.

At some point a restarted version of this study will likely focus on allogeneic use of IPSC perhaps via an IPSC bank being developed by Dr. Shinya Yamanaka. For many years the consensus, most exciting aspect of IPSCs in the field was considered to be their potential for use as the basis for powerful patient-specific autologous therapies. The apparent planned shift to non-autologous clinical use of IPSC in this case raises the question of how it would be superior or substantially different to the use of hESC, other than that making IPSC does not involve the use of a leftover IVF embryo.

This development also raises a 2nd question as to whether there will be a domino effect now of other clinical studies or trials that are in the works using IPSC switching to allogeneic paths as well. In other words, is this a historic, turning point moment for the IPSC field in Japan overall away from an autologous path? Or is the switch here to allogeneic just a one time, one study decision? More info on the regulatory changes is needed to help clarify the answer to this question and the path forward as well.

Hopefully the regulatory body in Japan (Ministry of Education?) that has made or is making the relevant regulatory changes will announce them publicly in detail soon.  If that information is already out there (e.g. in Japanese on the web) perhaps someone can find it and we’ll post it here.

24 thoughts on “Historic turning point for IPS cell field in Japan?

  1. It is inevitable for both hESCs and iPSCs to acquire mutations while being cultured- when those mutations give a selective advantage to the cultured cells, they will eventually dominate.

    Some of these mutations might be harmful if the cells are used in transplantation, but there have been no studies that ask what mutations are bad and which ones are benign in transplants. We really need to do this.

    It’s not possible to know that “the 2nd patient’s IPSC reportedly had six mutations that were not present in the original somatic cells”. Think about it: each iPSC clone comes from one somatic cell. Unless you did whole genome sequencing on that exact somatic cell, by itself (which you can’t do, because it’s the source of the iPSC clone), there is no way to know if the variants originated in the somatic cell or were acquired during expansion of the iPSC clone.

    WGS of a culture cannot detect minor variants in the population – we know that cultured cells are heterogeneous, but can’t measure that by sequencing. I would expect to see hundreds of genomic differences between the population of somatic cells and any iPSC clone derived from a single cell in that population.

    I repeat, we need to ask what mutations are benign and which convey risk.

    • Hi Jeanne,
      These are really good points.Thanks for bringing them up.
      Following up, how do we determine which mutations are benign and which are risk-associated in a transplantation context?
      Animal models?
      As to the source of the mutations, the New Scientist said: “Analysis of the patient’s cells revealed six mutations. Three were genes that had been deleted and three were changes to nucleotides, including one in an “oncogene” linked with a low risk of cancer. The mutations were not detectable in the original skin cells from the patient.” Presumably they got their information from Dr. Takahashi.
      Of course you are right that it is possible that the dermal fibroblast cultures from the 2nd patient collectively did not show detectable levels of these specific 6 mutations and yet they were present in the individual cells from which IPSC line(s) arose. One thing they could do is Sanger sequence say 100-1,000 individual cells from this culture at the loci of the mutations. While this would not be perfectly definitive it could be quite informative. Why not try to take advantage of this unfortunate situation to try to learn more about how such mutations arise? I would imagine that these cells were produced with minimal passaging and under ideal conditions to minimize risk of mutations arising.
      Another question is whether all 6 mutations are present in individual single cells or if they are scattered between heterogenous cells within the population.

  2. Hi Paul-

    It is clear that these mutations arise all the time, whenever cells divide or repair their DNA. Our observations are that variants are randomly distributed all over the genome in iPSCs.

    If we were to sequence a bit of all of your tissues (I promise not to do that!) we would find that you are mosaic–since the cells are under different selective pressures, you wouldn’t have the same range of mutations that are found in cultured cells, but you’d have a lot of detectable variants. Your cells would still be more similar to each other than to anyone else, but I’ll bet you would have hundreds of variations among the tissues.

    One experiment I’d like to do is to clone the fibroblasts- not allowing them to expand too much and become very heterogeneous, and compare the sequences of the clones to iPSCs generated from them. This would only be useful if the fibroblast clones were representative of the whole population, and that reprogramming doesn’t favor particular fibroblast genotypes.

    Yes, the only way to find out if the mutations are dangerous in vivo is to do a ton of transplants with different mutations. My opinion is that we could do a thousand transplants without finding one that was abnormal.

    I want to reiterate that both hESCs and iPSCs acquire mutations, and neither is better than the other. The advantage of an allogeneic cell therapy is that you could save a lot of work by making huge batches of cells and testing them thoroughly (don’t forget to check them after they are differentiated!). However, immunosuppression, I was told, costs $100,000 per person, so that, and the adverse effects of rejection, should be taken into account.

  3. The information of 6 genes’ mutations might come from a different source. I did not say anything about it.

    We chose RPE cells because we know that RPE never form tumors even they have gene mutations. There is no report about the metastatic tumors of RPE cells even in the familial cancer patients who have mutations in oncogenes such as P53. Furthermore, they produce PEDF that is a strong anti-tumor factor that kill the proliferating cells (we report it). Eye balls are filled with retinoic acid that will not keep immature cells. RPE is very special cells and the eye ball is special place both suitable for the first trial. We confirm fairly sensitive tumorigenicity tests (10000 times sensitive than one that use nude mice) again and again. So that as for RPE cells, gene mutations are not harmful as people think.

    Dr. Loring pointed out really important facts. We also think so.
    Since we knew that genetic analysis would cause this kind of debate and would not be solved for many years, there is a consensus in the committee in the ministry of education after more than 5 years discussion. Detailed genetic analysis is not required as the quality control “only” for RPE cells. Of course we did the WGS and other analysis as a research.

    Please distinguish the quality control and research.
    Confusing these two will leads very primitive discussion.

  4. The iPSC line bank will allow for haplotype matching and will cover I think something like 90% of the Japanese population. In this case the level of immunosuppression would be minimal and so would reduce the cost. However if we consider a country that has citizens of many ethnic backgrounds the corresponding iPSC bank would need to be pretty large. To get round this, we would need to share lines globally. For instance, american citizens of recent Japanese decent could use the iPSC lines being created at CiRA, Kyoto University. Similarly americans living in Japan as permanent residents could use lines now being created at the national labs in US.

    Whatever the source of cells I think ultimately we are going to need to screen everything before we transplant. Even after creating lines that are mutation free, we need to expand cells, differentiate cells and during these billions of cell divisions it is obvious mutations will creep in. Knowing what mutations are permissible will help greatly but I hope we can find a less anima intensive method, perhaps organ-on-chip will allow for rapid, high-throughput screening.

  5. I think I’m understanding more about the situation. Was the concern only with protein-coding regions? In other words, were the mutations discovered by exome sequencing? If so, I will revise what I said earlier. Although WGS reveals hundreds of variants (SNPs and indels) between parental fibroblasts and iPSCs, only a handful (1-7 in our study) were nonsynonymous or nonsense mutations in exons.


    • @Jeanne
      My impression was that these were CDS mutations, but perhaps Dr. Takahashi would clarify. Another question is whether all the mutations (assuming New Scientist got it right at 6) were in a single IPSC line from the 2nd patient.

  6. maybe a stupid question … but I think because of the procédure to produce IPS-cells the new IPS-cells have developed mutations. Why it is not possible to sort out the cells without mutations, to isolate them and to let these cells grow again? These new grown cells will likely not develop the new mutations, because they won`t need an additional procédure to turn in IPS-cells?

    • Not a stupid question. The mutations arise when the cells divide, so they would get worse as cells are cultured. The reason they show up is selection of the fittest. The cells that resist apoptosis or divide faster in a culture dish will dominate the population after the cells are expanded. Stem cells also avoid apoptosis and divide faster, but these traits are much much MUCH stronger in cancer cells.

  7. @Jeanne – I also have a possibly stupid question – why would iPSC, which have been moved out of a post-mitotic state into an immortal state, be under the same selection pressure to mutate growth suppressive genes as primary cells that would otherwise hit crisis. Aren’t stem cells past the need for mutating growth arresting genes and are hence unlikely to collect tumorigenic changes?

    • It’s a fine question. The fibroblast cells that we use are not post-mitotic, but they are contact inhibited. We have to culture them for quite a while to have enough for reprogramming.

      About the growth arresting genes, I guess that pluripotent stem cells violate that basic principle about immortalized cells…we know that some hPSC lines grow faster than others, and the ones that grow faster have the common aneuploidies and duplications that have been reported by several labs…Chromosome 12, 17, and X aneuploidies, and duplications on Ch20 are the most common. None of those has been reported to be tumorigenic.

  8. Why not just use up and coming targeted cancer therapies in culture to put negative selection pressure on the derived cells with cancer like mutations that cause them to out-compete the regular cells in vitro? Or even non invasive sorting of cancer like cells?


    Apologies if this is a layman question about pie in the sky goals that have already been considered elsewhere ad nauseum.

  9. @jeanne – many thanks – I forgot the fibroblast was still a humble fibroblast!

    Interesting about hPSC accumulating abnormalities. For the generation of hPSC clones for clinical application, I guess we’re not going to avoid cell replication so we need to thoroughly screen clones for key mutations, as the Takahashi lab did, or somehow select against mutations as Jim Bacon suggests. Takahashi’s method seems the most cost-effective.

  10. Jim:

    Interesting idea. I think that if there were cancer cells in the population they might be “softer” as the UCLA researchers reported. But I’m not sure how an atomic force microscope could be applied to practical sorting of cells.

    My guess is that we couldn’t detect any cancer cells in the hPSC populations, because so far in everyone’s detailed genomic analysis, no one has found true cancer cells in hPSC cultures. (anyone out there, correct me if I”m wrong).

    A thought experiment: It would be interesting to mix marked hPSCs with real cancer cells, and see what if any cells show this squishy cancer phenotype when they are under the same conditions. But because cultured cells are not like cells in fluids in the body, I would predict that it wouldn’t be simple.


    I agree. Even whole genome sequencing is relatively inexpensive these days, and SNP genotyping is cheap. There’s no excuse for not checking the cells for mutations. We should keep multiple cultures of the same cell line, so we can toss out the losers.

  11. Hello, I just have a question as a non scientist:

    Does this method from Dr Kosai ( 21 july) could resolved some issues of tumor cells risks? Or is it some “too optimistic claims”?


    “The researchers created strains of a genetically modified virus that can proliferate and kill cells that contain survivin. Seven days after the modified virus was introduced into undifferentiated iPS cells and ES cells, all those cells died.
    Meanwhile, cells that had turned into the specific tissue survived the introduction of the virus.
    The research team will present its findings at a conference of the Japan Society of Gene Therapy on July 25 in Osaka.”

    Ken-ichiro Kosai (Department of Gene Therapy and Regenerative Medicine / Center for Innovative Therapy Research and Application, Kagoshima University Graduate School of Medical and Dental Sciences)

    “Development of the original survivin-responsive conditionally replicating adenovirus toward the investigator-initiated GCP clinical trial”

    Does this could help to expand human clinical trials regarding regenerative therapies?

    Best regards,

  12. @Richie – I also have diabetes mellitus and I more than halved the amount of insulin I needed and brought my blood sugar control back to normal. I also had “significant increases in energy levels, a marked decrease in fatigue, improved muscle tone and pains in my legs disappeared” …and so on, like the patient in the report. Did I get stem cells too? No – I got a training bike and cut out bread and pasta from my diet.

    These reports are wantonly misleading with no credible science or medicine to back them up. They give no indication on how the cells are supposed to work – because they cannot prove it. And if stem cells can somehow “cure” diabetes, why does it even exist when we all have billions of stem cells already? The only difference is that some have a $30,000 price tag – a bike is much cheaper.

  13. @Muggles,
    Every time I read about new possible stem cell applications I am both enthusiastic and sceptic. There are already good stem cell therapies for example in the orthopedics. And of course there are also many offers without any evidence, so the difficult challenge is to judge a maybe new therapy.

  14. In addition to numerous flaws of iPSC reprogramming which is conveniently disregarded of the workers in this filed, one has to realize that during of iPSC generation close to 80 genes including oncogenes and IGFR loose their genomic imprinting. In this sense, the hESCs are much superior in lacking major changes in imprinting. I am worried that even the first patient will develop tumors over time. Question is how this will be detected?

    Also very curious is why this information about halting the clinical trial was not announced at ISSCR. The information about halting the clinical trial was made public two weeks after the meeting. There is no question in my mind that this information was purposefully released after ISSCR to shelter this high risk ill-advised clinical project.

    If Dr. Takahashi knew at the time of ISSCR meeting that this clinical trial was going to be halted and this information was not released then this should be investigated. By not doing so her group and organization has mislead thousands of students and scientists.

    • @Michael,
      From what I understand, sometimes there can be many complexities to clinical studies, regulations, and such that influence or even control when one might be able to talk about a finding or situation in the public domain. I would encourage care in the choice of one’s words and let’s all learn more over time before making accusations.

  15. Michael-
    I don’t know where you got this idea, but iPSCs and hESCs both have aberrant imprinting- iPSCs no worse than hESCs. You should read my group’s paper on DNA methylation of iPSCs and hESCs – it has a lot of information on changes in DNA methylation as hPSCs are cultured and differentiated. Cell Stem Cell 2012.

  16. Thank you all for the informative comments.

    I have a simple question, which part of the process poses the highest risk of developing tumorigenic mutations? is it the expansion of patient’s fibroblasts before reprogramming? or is it the expansion of generated iPS lines before differentiation?. The expansion of iPSC’s for RPE generation might be minimal when compared to generating enough cells for cardiac or pancreatic treatments, where ~2×10^9 cells are thought to be required.

    Seeing the struggle with the safety of RPE cell sheet makes one wonder how far are we from seeing iPS-based therapies for heart failure or diabetes.

Comments are closed.