CRISPR Digest #14

Two years ago, in spring 2015, Liang et al. published the first report of gene-editing in human embryos using CRISPR/Cas9 (mentioned previously here, here and here). At the time no high-profile journal was willing to take on the risk of publishing what was perceived to be a controversial study. Liang et al. were trying to correct mutations in the human beta-globin gene – mutations in this gene can lead to a group of diseases called beta thalassaemias, including sickle cell anaemia – in human embryos that had been fertilised by two sperm cells (and could therefore never develop). In fact, the take-home message from their study was that using the techniques available to them at the time led to a host of unwanted side effects, including the creation of mutations at other sites in the embryo genome and the “correction” of the beta-globin gene with a similar gene called delta-globin.

Last month, a different group (Ma et al. – four first authors and five corresponding authors!) published more work on human embryo CRISPR/Cas9 gene-editing, this time in Nature. Like Liang et al. this paper also tried to tackle a monogenic disease, a disease that is caused by a well-defined mutation in a single gene, called hypertrophic cardiomyopathy. The affected gene is MYBPC3 and when mutated (denoted as DeltaGAGT in the figure below) this leads to a thickening of the heart muscle, which in turn can cause heart failure. The authors used donor sperm with the MYBPC3 mutation together with healthy oocytes to perform their experiments. In the first approach the eggs were fertilised by the sperm and only subsequently, during S phase, were the guide RNA, Cas9 protein and a piece of non-mutated donor DNA injected. The guide RNA was designed to specifically recognise the mutant version of MYBPC3, which recruits the Cas9 protein to make a cut in the DNA, and then the donor DNA would serve as a template to repair the sperm’s mutated gene. Ma et al. observed that this technique worked but often generated so-called mosaic embryos, which contained a mixture of healthy and mutated cells. This incomplete gene correction happened because during S phase both the maternal and paternal chromosomes duplicate and therefore the CRISPR/Cas9 system would have to correct two mutated MYBPC3 genes before the first cell division.

Screen Shot 2017-09-10 at 16.16.44
Schematic depicting CRISPR/Cas9 stage at zygote stage (top) versus together with sperm (bottom) – copied directly from Ma et al, 2017

In a second approach, Ma et al. wanted to overcome this mosaicism by injecting the CRISPR components together with the sperm during the M phase of the oocyte. Now only one copy of mutant MYBPC3 had to be corrected and this succeeded in producing completely healthy embryos. Ma et al. also checked to make sure that these embryos did not carry any unwanted, off-target mutations.

Last but not least, Ma et al. provided evidence that often the human zygote used the healthy maternal gene to provide a template for the repair of the mutated paternal gene, instead of the injected DNA template. This is significant because in most cell types the DNA double-strand breaks caused by Cas9 are usually repaired in an imprecise manner (called non-homologous end joining) and lead to further mutations. Ma et al. therefore argued that “human gametes and embryos employ a different DNA damage response system”.

This finding could be of huge importance, both to the basic understanding of human embryonic development as well as to potential therapeutic CRISPR/Cas9 applications. However, four days after the Nature paper was published online, several prominent scientists posted a riposte on the pre-print server bioRxiv. Egli et al. criticised the first paper quite heavily by raising theoretical objections/concerns; they couldn’t have tried to replicate the experiments in such a short time frame. [Note that this pre-print was, of course, not peer-reviewed, although the authors have confirmed that they were trying to get their work published in Nature as well.]

Among other more technical issues to do with the way in which healthy and mutant genes were detected, Egli et al. pointed out that after fertilisation the maternal and paternal chromosomes remain physically separated (indicated by the arrows in the figure below) until just before the first cell division. Therefore, Egli et al. argued, it is highly unlikely that the healthy maternal MYBPC3 gene could serve as a template for the repair of the mutant paternal gene. This strikes me as a strong argument, not being at all familiar with early human development. Overall, Egli et al. suggested that Ma et al. were simply not detecting the mutant gene in their embryos but not providing good enough evidence of a corrected gene. The scientific debate will, no doubt, continue and I think having bioRxiv as such a rapid place for the exchange of ideas can drive scientific discourse.

Egli et al - early development
Pictures of a human zygote (fertilised egg/oocyte) and its very early development – copied directly from Egli et al, 2017

Since this is a digest it should also contain some other relevant CRISPR/Cas9-related news. One of the post docs I met at Cold Spring Harbor Laboratory in 2014, Serif Senturk, published a paper early this year in which the authors show how they can switch CRISPR on or off in living cells. They did this by fusing the Cas9 protein to another, destabilising protein domain, which caused the attached Cas9 to get degraded. However, when a “shield molecule” was added to the cells, the destabilising domain was no longer active and the Cas9 could accumulate. This innovation counteracts the problem of off-target effects, which are often due to the long duration that Cas9 is active for. Pretty neat system, I think.

Senturk 2017

Schematic depicting Cas9 fused to a destabilising domain – copied directly from Senturk et al, 2017


Egli D, Zuccaro M, Kosicki M, Church G, Bradley A, Jasin M (2017) Inter-homologue repair in fertilized human eggs?

Liang P, Xu Y, Zhang X, Ding C, Huang R, Zhang Z, Lv J, Xie X, Chen Y, Li Y, Sun Y, Bai Y, Songyang Z, Ma W, Zhou C, Huang J (2015) CRISPR/Cas9-mediated gene editing in human tripronuclear zygotes. Protein Cell: 1-10

Ma H, Marti-Gutierrez N, Park SW, Wu J, Lee Y, Suzuki K, Koski A, Ji D, Hayama T, Ahmed R, Darby H, Van Dyken C, Li Y, Kang E, Park AR, Kim D, Kim ST, Gong J, Gu Y, Xu X et al. (2017) Correction of a pathogenic gene mutation in human embryos. Nature 548: 413-419

Senturk S, Shirole NH, Nowak DG, Corbo V, Pal D, Vaughan A, Tuveson DA, Trotman LC, Kinney JB, Sordella R (2017) Rapid and tunable method to temporally control gene editing based on conditional Cas9 stabilization. Nature Communications 8: 14370


Combining Art and Science

This is a confession. I love cells. However, if this were an essay for my German teacher at secondary school she would, given the name of this blog, award me a “Themenverfehlung” (basically completely missing the point of the essay question), which she did on a semi-regular basis. I could argue that in the broadest sense this post is also about cells, because cells were necessary for the creation of this work of art as well as for the analysis. But I admit that’s a bit of a stretch.

Anyway. I found this new paper published in Scientific Reports (Thurrowgood et al, 2016), which is simply entitled, ‘A Hidden Portrait by Edgar Degas’. The Australian research group, comprised of art curators and X-ray technology specialists, used a technique called X-ray fluorescence to study “Portrait of a Woman”, which hangs in the National Gallery of Victoria, Australia. Normal X-ray analysis had already revealed that an earlier portrait of a different woman lies underneath what is currently visible:

figure 1

Visible image of “Portrait of a Woman” (left) and conventional X-ray imaging (right) – copied from Figure 1

Conventional X-ray technology relies on the different densities of the materials/pigments in the painting. The denser areas absorb more of the X-rays and the X-ray film behind is less exposed than behind less dense areas, just like in a chest X-ray, for example. However, X-ray fluorescence is a more sensitive technique because it can detect specific elements in the painting. When atoms are hit by high energy X-rays some of the electrons close to the atomic nucleus (the “inner shells”) can be flung out of the atom entirely. This leaves a hole for electrons to fill and when these electrons “fall” into the inner shells energy is released in the form of electromagnetic waves, which are characteristic for different atoms:


Schematic of X-ray fluorescence – copied from this link

This technique is not entirely new, but as far as I understand this was the first time it has been used for a large-scale analysis of a painting. Using the physical properties of different atoms, Thurrowgood et al. created so-called elemental maps to determine which parts of the hidden portrait contained which elements.

figure 3

Elemental maps – copied from Figure 3

By analysing which elements are found in the same areas the researchers could trace back which pigments were used and therefore make educated guesses about the original colour of the painting. For example, the hair contained a lot of iron and manganese, which are often found together in the brown umber pigment.

figure 6

False colour image reconstruction of the hidden portrait of Mlle. Dobigny – copied from Figure 6

Mademoiselle Dobigny was one of Degas’ favourite models, but who knows why he kept her portrait for so long before deciding to paint over it. The reasons are certainly less obvious than in Georges Seurat’s painting “Young Woman Powdering Herself” (Courtauld Institute of Art, London): he concealed the only known self-portrait of his behind a mirror, watching the young woman who also happened to be his lover…

Reference and further reading:

Thurrowgood D, Paterson D, de Jonge MD, Kirkham R, Thurrowgood S, Howard DL (2016) A Hidden Portrait by Edgar Degas. Scientific Reports 6: 29594

New York Times article on Degas’ painting

Courtauld Institute of Art blog on Seurat’s painting

Editing @eLife

Within the span of three days two people independently told me that I “always manage to land on my feet”. Well, I think they’re right about that. For the last two weeks I have been working as an editorial intern in the Features Team at the scientific journal eLife.

eLife logo

eLife logo (copied directly from their website)

They won’t object to me re-using their logo here since eLife was established in 2012 as one of the first fully open-access online journals for the life and biomedical sciences. One of eLife’s distinguishing features is its fast decision-making process: once a research team submits an article for consideration the initial decision of whether to review it or immediately reject it is made within a few days. Furthermore, the reviewing process itself rarely takes more than a month and if changes need to be made to the manuscript then there usually is only one round of revisions. So when it took them a whole nine days to reject a paper I had been working on with friends they were actually being slow. Other than having a fast decision-making process, eLife is also trying to increase the quality of the science they publish and one endeavour that I find particularly interesting is their “cancer reproducibility project“: 50 of the highest impact cancer papers published between 2010 and 2012 were picked and are now going to be re-done by independent researchers in an attempt to find out how reproducible the results actually are.

In the day-to-day publishing at eLife, however, one of the main things I worked on was writing so-called Digests – short summaries of each of the research papers that are meant to be understandable to interested laypeople. The digests include some background information, an explanation of the main results in the paper and a brief description of which questions future experiments will address. Among other things, I wrote about what happens to sleep-deprived fruit flies, a new mechanism that protects against pancreatic cancer, how some pathogenic gut bacteria get past our defences, and how skin cancer cells move. After reading and thinking about a considerable number of these articles it is less surprising that our paper was rejected.

Of course my digests are not being published as I wrote them. Peter, the main Features editor, went through and corrected all of my writing, something that was extremely useful to me. For instance, I learnt to avoid what he calls “jaw-breakers”, combinations of words in quick succession that are difficult to say. (However, upon re-reading that last sentence, maybe I should say I am still “learning to avoid” jaw-breakers.) Other things he pointed out were practical, because they helped me understand how much or how little the general public can be assumed to know about science. By the start of the second week I could already hear Peter’s voice in my head while writing – which sounds a lot worse than it was since he actually has a pleasant Northern Irish accent – telling me to rephrase this or shorten that.

Apart from writing digests I edited a so-called Insight article, a slightly longer article that comments in more depth on an original research paper and is written by an expert in the field. In particular, I learnt about how the production of transgenic pigs might be able to curb the next outbreak of foot-and-mouth disease (assuming these pigs will be approved by the various regulatory bodies).

Other than that I worked on editing an interview with an early career researcher, i.e. a researcher who is at PhD level or higher but has not received a tenure-track position (yet). In between these things I proofread some articles before they were published in their final form, or looked for “pull quotes” to make articles more interesting. (I didn’t know that pull quotes were called that, but they are the sentences that are pulled out of the text of an article and enlarged so that you immediately read them and subsequently get slightly annoyed when you find them again in the main text.)

the features team

Stuart, Peter, me, Emma and Sarah (from left to right)

I have to admit that working in a typical open-plan office was a new experience for me, since I’m generally used to chaotic lab benches and cramped desk spaces. The thing I enjoyed least was that everyone eats lunch on their own (except for Friday pub lunches) and often at their desk. This struck me as quite strange since you would expect it to be much easier to coordinate having lunch together in a place where everyone works at a computer. In labs people are busy doing experiments but somehow they still find time to be a bit more sociable during their breaks. However, having a nine-to-five (or 9.30 to 4.30 …) job is certainly one of the perks of working in an office like this. Furthermore, I think there is generally less chance of one taking work home (both physically and mentally), although of course I imagine that changes as one assumes more responsibility.

Overall I had an enjoyable experience and am grateful for having been able to get a glimpse into real scientific writing, editing and publishing. If I continue writing eLife digests as a freelancer this will give me the benefit of keeping up to date with the latest, high-quality biomedical research outside the narrow range of a PhD. So really all there is left to say is Thank You to Peter, Emma, Sarah and Stuart, and the rest of the eLife team for being so helpful and welcoming.

More Stem Cell News

Since the last time I wrote about stem cells, several things of note have occurred. Among these was the announcement that the Japanese researcher Haruko Obokata has resigned from the RIKEN Center for Developmental Biology, because the finding from her paper (Obokata et al. (2014)) – that she could induce a stem cell state by treating differentiated cells with acid – could not be reproduced. An investigation by the RIKEN Center has found that the supposedly reprogrammed cell cultures were likely contaminated by embryonic stem cell lines, but it is still unclear whether that was accidental or deliberate.

On a more positive note (assuming that these results are true), Irie et al. (2014) have been able to demonstrate the reprogramming of differentiated human epithelial cells into induced pluripotent stem cells (iPSC), which they then re-differentiated into primordial germ cells (PGC), the precursors of germ cells (egg and sperm cells). In particular, they found that the transcription factor SOX17 is a main regulator of this cell state in humans, whereas it is a different transcription factor, BLIMP1, in mice. The implications of this work are potentially vast (see the Nature blog comment here). For instance, when the PGCs generated from differentiated mouse cells were transplanted into mouse testes or ovaries they developed normally into sperm or egg cells, respectively, and were then amenable to in vitro fertilisation. Potentially, therefore, this recent advancement in human cells may be a step towards treating infertility, or enabling (male) homosexual couples to have their own biological children. However, there are biological, technical and ethical hurdles to be overcome if these findings are to be applied. The following figure summarises their findings and is copied directly from their graphical abstract:

Irie et al., Graphical abstract-final5_wt

The full text of Irie et al. can be viewed here as an open access text. And this is because the research and publication costs were covered by the Wellcome Trust, a “global charitable foundation dedicated to achieving extraordinary improvements in human and animal health”, which supports “public engagement, education and the application of research to improve health”. Many funding bodies, including the Wellcome Trust, are increasingly imposing the restriction that any of the research funded with their money must be published open access, including the sharing of raw data.

As one would say in German, I wish you “a good slip into the New Year”!


Irie N, Weinberger L, Tang WW, Kobayashi T, Viukov S, Manor YS, Dietmann S, Hanna JH, Surani MA (2014) SOX17 Is a Critical Specifier of Human Primordial Germ Cell Fate. Cell

Obokata H, Wakayama T, Sasai Y, Kojima K, Vacanti MP, Niwa H, Yamato M, Vacanti CA (2014) Stimulus-triggered fate conversion of somatic cells into pluripotency. Nature 505: 641-647

#OA = Open Access



This is the symbol for open access (image from PLoS, the Public Library of Science). Two days ago Nature Communications, a “journal that publishes high-quality research from all areas of the natural sciences” within the larger and prestigious Nature Publishing Group, announced that it will only be accepting open access (OA) submissions starting next month. Yesterday the Royal Society in London tweeted to announce the launch of their new OA journal Royal Society Open Science.

So what exactly is OA publishing? Simply put, it is the online publishing of research results that does not put any restrictions on who can read the articles, and additionally, allows further distribution and re-use of the material as long as the original source is cited. This relatively new model of publishing stands in contrast to the well-established traditional publishing model, in which readers can only access journal contents when they (or the institutions they work for) pay subscription fees. Generally, there are two types of OA publishing (for more information and resources I found the site of Peter Suber very helpful):

  1. “gold OA”: in the form of OA journals in which the papers are peer-reviewed before publication; examples include PLoS and BioMed Central (for a full list see the Directory of OA Journals)
  2. “green OA”: in the form of OA repositories in which articles are stored online and may or may not have gone through a peer-review process

If the reader no longer pays for access to information, who provides the money for the editing and publishing process? There are several models to finance OA publishing. A common model has the authors (or the authors’ institutions) pay for their publication once the paper has been accepted; this usually occurs in the form of an Article Processing Charge (APC). These fees are far from negligible: Royal Society Open Science charges £1000, PLoS Biology approximately £1800 and Nature Communications a whopping £3150.

The first drawback of OA immediately becomes obvious: not all researchers have a budget large enough to pay for the publication of their work after having spent a lot of money on the research itself. Does this mean that OA is biased towards making big and successful labs even more successful? Thankfully, OA journals usually offer to waiver the APC if the research group is based in a low-income country, and often large funding bodies support their scientists by paying for the APC as well. A way to think of the concept is to liken it to social security: those that can pay for their publications, and in doing so also support those that cannot afford the APC as well as providing money for the review process of those manuscripts that are eventually rejected.

Another concern that arises is so-called “predatory OA”, in which publishers try to make a profit from the APC without rigorously reviewing and editing papers, thus increasing the amount of badly curated information available on the internet. To counteract this trend Jeffrey Beall regularly updates a list of journals that are “questionable”.

Lastly, the question arises of whether more people actually read journal articles just because they can. Do laypeople read about discoveries on diseases in PLoS Pathogens or about replication studies and negative results in PLoS One? I don’t know the answer to this, but probably not. OA is partly an ideological step forward: who pays for most of the research conducted in institutions and universities? Taxpayers. So who should be allowed to read the outcome of this research? Taxpayers.

Assuming that laypeople do not benefit from OA, who does? Scientists/academics at small institutions that cannot afford to buy subscriptions to all journals; recent graduates that want to keep up with current scientific literature but no longer have access via their institutions; doctors with private practices that are not affiliated with hospitals: only a few days ago my mother tried to find information on the treatment of a complicated case of endometriosis (a disease in which cells of the uterine lining grow outside the uterine cavity), but she couldn’t access papers published in The Lancet or The New England Journal of Medicine, so I had to help her out by signing in using the university’s subscription.

Another perk of online OA publishing is that it is less restrictive with regards to the length of the papers published. Unlike in journals such as Science, where high impact research is squeezed on to three pages (with dozens of pages of supplementary material), OA journals can afford to “print” long articles enabling scientists to show off all their hard work.

What will the future trend be? Will all journals become OA or will there continue to be subscription-only publishing in science? From what I have read and experienced there is certainly a tendency towards more OA, also among the established high impact factor journals in the form of green OA (making the paper available six months after initial publication). Additionally, there are archives that make papers available before they are printed (e.g. arXiv for physics and computational sciences or bioRxiv for biology). It remains to be seen what long-term benefits OA publishing will have.