A Wellcome Trust/Royal Society funded Research Associate position is available in Dr. Kyra Campbell’s research group. This is a fantastic opportunity to join the Campbell group, who are focused on identifying the molecular mechanisms underlying epithelial cell plasticity during development and disease. We study this during morphogenesis of the Drosophila midgut (Campbell et al, Dev Cell 2011; Campbell and Casanova, Nat Comms 2015), and also in exciting Drosophila cancer models that we have recently generated (Campbell and Casanova, Plos Genetics 2018; Campbell et al, Nat Comms 2019).
We are combining single-cell OMICs approaches and deep-tissue imaging on our own labs dedicated multiphoton confocal microscope, with genetic approaches and CRISPR/Cas9 technologies. We are looking for a motivated and enthusiastic candidate who will play a central role in the lab. You must have a good honours degree and a PhD (or be close to completion) in areas relevant to cell/developmental biology (or have equivalent experience), along with experience in in vivo imaging and image analysis. Applicants are expected to have excellent interpersonal and communication skills, be highly independent and committed to research in a fast-moving and competitive field.
All cells in the body contain the same genetic material. The difference between cells therefore depends solely on which genes are expressed or ‘turned on’. Now, researchers from the University of Copenhagen have gained new insights into how genes are turned on and off and how the cells “forget their past” while developing into a specific cell in the body. This new knowledge is published in Nature and will be crucial for stem cell therapy and potentially treating people with cancer.
Stem cells all share the potential of developing into any specific cell in the body. Many researchers are therefore trying to answer the fundamental questions of what determines the cells’ developmental fate as well as when and why the cells lose the potential of developing into any cell.
Now, researchers from the Novo Nordisk Foundation Center for Stem Cell Biology (DanStem) at University of Copenhagen have discovered how stem cells can lose this potential and thus can be said to “forget their past”. It turns out that the proteins called transcription factors play another role than the scientists thought. For 30 years, the dogma has been that transcription factors are the engines of gene expression, triggering these changes by switching the genes on and off. However, new research results published in Nature reveal something quite different.
“We previously thought that transcription factors drive the process that determines whether a gene is expressed and subsequently translated into the corresponding protein. Our new results show that transcription factors may be more analogous to being the memory of the cell. As long as the transcription factors are connected to a gene, the gene can be read (turned on), but the external signals received by the cells seem to determine whether the gene is turned on or off. As soon as the transcription factors are gone, the cells can no longer return to their point of origin,” explains Josh Brickman, Professor and Group Leader, DanStem, University of Copenhagen.
The question of how a cell slowly develops from one state to another is key to understanding cell behavior in multicellular organisms. Stem cell researchers consider this vital, which is why they are constantly trying to refine techniques to develop the human body’s most basic cells into various specific types of cells that can be used, for example, to regenerate damaged tissue. So far, however, investigating the signals required to make cells switch identity has been extremely difficult, since making all the cells in a dish do the same thing at the same time is very difficult.
A protein centered viewpoint
The researchers developed a stem cell model to mimic a cell’s response to signaling and used it to, for first time, precisely determine the sequence of the events involved in a gene being turned on and off in response to a signal in stem cells. The researchers were able to describe how genes are turned on and off and under what circumstances a cell can develop in a certain direction but then elect to return to the starting-point. Part of this work involved measuring how proteins in a cell are modified by phosphorylation using advanced mass spectrometry available through an important collaboration with Jesper Olsen’s Group at the Novo Nordisk Foundation Center for Protein Research. “Combining forces with the Olsen group in the CPR enabled us to provide a unique deep description of how individual proteins in a cell react to signals from the outside,” continues Josh Brickman.
New answers to old scientific questions
These results are surprising. Although the sequence of cell transcription processes could not previously be measured as accurately as in this study, the dogma was that transcription factors comprise the on-off switch that is essential to initiate transcription of the individual gene. This is not so for embryonic stem cells and potentially for other cell types.
“Transcription factors are still a key signal, but they do not drive the process, as previously thought. Once they are there, the gene can be read, and they remain in place for a while after the gene is read. And when they are gone, the window in which the gene can be read can be closed again. You can compare it with the vapour trails you see in the sky when an airplane has passed. They linger for a while but slowly dissipate again,” explains first author, William Hamilton.
This discovery is first and foremost basic knowledge, which changes fundamental assumptions in molecular biology. The new results are especially important for researchers working on stem cells and cancer biology. They provide new insight into how cells develop, how pathways involved in development determine when cells change, and when the point of no return is reached. These pathways are also found frequently mutated in cancer and the findings in this study will be valuable to the study of malignant development.
“In the project, we focused on the fibroblast growth factor (FGF)–extracellular signal–regulated kinase (ERK) signalling pathway, which is a signalling pathway from a receptor on the surface of a cell to DNA inside the cell nucleus. This pathway is dysregulated in many types of cancer, and we therefore hope that many of the data in this study will help to inform aspects of cancer biology by indicating new ways to specifically target this signalling pathway in cancer cells,” concludes Josh Brickman.
They study was funded by the Novo Nordisk Foundation, the Independent Research Fund Denmark, the Danish National Research Foundation, the Human Frontier Science Program and the Lundbeck Foundation. It also involved an important collaboration with the group of Naama Barkai, at the Weizmann Institute for Science, Rehovot, Israel.
Novo Nordisk Foundation Center for Stem Cell Biology, DanStem, University of Copenhagen | joshua.brickman@sund.ku.dk
Josh Brickman has a background in molecular biology and gene regulation. From a PhD focused on transcriptional regulation he trained in developmental biology as a post-doctoral fellow, working in early mouse, and Xenopus, as well as cultivating embryonic stem cells as a model for developmental biology. He began his own lab with research projects bridging early development in multiple models systems with ES cells in a hybrid approach aimed at understanding conserved mechanisms of lineage specification, pluripotency and self-renewal. He currently seeks to understand how transcription factors regulate cell fate choice in ES cells and early embryos. More specifically, Professor Brickman’s and his group investigate the basis for transcriptional priming and commitment in ES cells and early in the specification of the endoderm lineage. They hope to understand the relevance of these molecular events to cellular decision making, pattern formation, in addition to stem and progenitor cell potency.
Novo Nordisk Foundation Center for Stem Cell Biology, DanStem, University of Copenhagen | william.hamilton@sund.ku.dk
William obtained his PhD at the Edinburgh University in the labs of Tilo Kunath and Mike Tyers, where he worked on defining factors that regulate MAPK signalling in mouse embryonic stem cells. He then joined the Brickman lab in Copenhagen where he expanded upon this to uncover how MAPK signalling regulates transcription and plasticity during early stem cell differentiation.
Drosophila wing discs are epithelial sac-like organs and a powerful model for investigating the link between proliferation and patterning. Of particular interest is the question of how single cells in the disc integrate information regarding position and growth control, as morphogens that pattern an axis can also regulate cell division. A new Techniques and Resources article in Development reports the application of single cell sequencing technologies to dissociated discs in an effort to understand these problems. We caught up with first author Mingxi Deng and his supervisor Yan Yan, Assistant Professor at The Hong Kong University of Science and Technology (HKUST), to hear more about the story.
Mingxi and Yan (L-R)
Yan, can you give us your scientific biography and the questions your lab is trying to answer?
YY My lab is primarily interested in organ size control, in particular, the roles of cell structural components such as apicobasal polarity proteins and cytoskeletal proteins in this process. When I was a graduate student with Prof. Trudi Schupbach at Princeton University, New Jersey, USA, I performed a genetic screen for mutants affecting Drosophila follicle cell epithelial morphogenesis and proliferation. I then got my postdoc training with Prof. Chris Doe at the University of Oregon, USA, where I learned how Drosophila embryonic neuroblasts lose their apical domains and emerge from neuroepithelia. These experiences were important for me to learn that cell polarity and, more broadly, cell structural proteins, are important for organ size and shape. Another important thing I learned from graduate school is the power of quantification, which influences how we approach questions now in the lab.
Mingxi, how did you come to work in the Yan lab, and what drives your research today?
MD When I was looking for a postgraduate student position at HKUST I found Prof. Yan’s research interesting in combining the power of Drosophila genetics with quantitative biology methods. Her lab also has a good reputation for being supportive to students, so I decided to join. I have always wanted to become a scientist and I get excited when I encounter new problems and need to find a way to solve them.
What was the drive behind doing a single cell analysis of the disc, and how easy was it to set the system up?
MD & YY It started with our study of how scribble (scrib) mutant tumours – which show disrupted tissue architecture – change over time. We found that they showed a high degree of plasticity, and suspected that they might be more heterogeneous than previously assumed. For this we needed to understand how much of the cell heterogeneity in the scrib tumours comes from heterogeneity already existing in wild-type wing discs. That was the starting point of this analysis.
We were lucky to have the help we needed for this study. Prof. Ting Xie from the Stowers Institute for Medical Research, Missouri, USA, happened to visit my university at the time. I went to talk with him and he kindly shared his unpublished fly single cell dissociation protocol. It is also very helpful to have colleagues Jiguang Wang and Hao Ge with whom to discuss methods: their expertise in bioinformatics and mathematics ensures that we are analysing data correctly and robustly. In addition, the single cell community has been very good in providing open-access analytical tools with user-friendly tutorials.
Can you give us the most surprising finding from your paper?
MD & YY The most surprising finding is that pattern formation partially persisted in the scrib mutant tumours. This is surprising because the morphogens important for pattern formation need to properly spread in space, and it suggests that further studies are needed to understand why particular pattern formation processes are robust against loss of tissue architecture.
A combination of different gene expression patterns from the ‘Virtual wing disc in situ’ section of the online database.
What does your scrib tumour model analysis tell you about the link between patterning and growth control in the disc?
MD & YY This is a very good question, but we still do not understand the link between patterning and growth control, although we are able to make more quantitative observations from these data. It was previously shown, with very clear genetic evidence, that patterning factors such as dpp are needed to ensure proper proliferation and growth in the disc. Previous experiments have also shown that cell proliferation and growth are more or less uniform throughout the whole disc. Now, we provide another line of evidence that proliferation and growth states do not appear to be particularly biased in sub-regions marked by any single patterning gene, in both wild-type and scrib mutant discs. Our data also suggest that a well-defined distribution of proliferation and growth states exists in discs and this distribution is severely disrupted in the scrib mutants. Interestingly, the temporal scrib mutant data suggested a positive correlation between formation of correct patterns and a distribution of proliferation and growth states closer to wild type.
Your single cell datasets are available to explore on a database: what questions do you think this database will be particularly useful for addressing?
MD & YY Wing discs have been a very good system to study pattern formation, organ size control and regeneration. I hope that our database can provide a good reference point for the community interested in these questions. For example, for researchers interested in how wing disc cells respond to injury during regeneration processes, they would be able to compare the identity of their cells of interest with the wild-type imaginal disc cells in our database.
I hope that our database can provide a good reference point for the community
When doing the research, did you have any particular result or eureka moment that has stuck with you?
MD After I assigned the disc cells correctly to the pouch/hinge region, I was very happy to see that fine patterning processes are well represented in our single cell data. This gave me a sense of connection and also deep respect for the classical works on pattern formation, which I had previously only learned from textbooks.
And what about the flipside: any moments of frustration or despair?
MD I started as the only student working on computation in our lab and needed to learn everything from zero. My lab mates are all excellent experimentalists but they cannot help me with computational problems. It took a while to grow out of the loneliness but I have become more confident now.
So what next for you after this paper?
MD I have just finished my second year as a postgraduate student. I am now pursuing a few quantitative biology projects for which we already have data and a priority for me is to further sharpen my computational and mathematical skills. Hopefully, I can share new exciting stories in a few years when I graduate with my PhD.
Where will this work take the Yan lab?
YY The scrib mutant cells are very interesting, because when they are generated as mosaic clones in the wing discs they behave very differently and undergo cell death through a cell competition process. Building upon this work, we are now trying to better understand the scrib mutant clonal cells, and how different signalling activities contribute to their cell plasticity at the single cell level and eventually alter their growth outcome.
Finally, let’s move outside the lab – what do you like to do in your spare time in Hong Kong?
YY: I have a 4-year-old son and I am expecting another baby in December, so my activities outside the lab revolve around parenting. I find the parenting experience extremely helpful in that I am much more patient with students now than before.
MD: Hong Kong is a surprisingly great place for outdoor activities like hiking and sailing, which I like. I also like to play soccer and computer games.
In the latest Genetics Unzipped podcast we’re reporting back from the Manova Global Health Summit in Minneapolis last month, exploring the latest advances in health technology such as CRISPR-based gene therapies, infection-fighting bacteriophage and the possibility of curing HIV with stem cell transplants.
Plus veteran New York Times columnist Jane Brody’s advice for a healthy life, and reflections on progress in cancer from US journalist and advocate Katie Couric.
If you enjoy the show, please do rate and review and spread the word. And you can always send feedback and suggestions for future episodes and guests to podcast@geneticsunzipped.com Follow us on Twitter – @geneticsunzip
We study stem cells, development, and regeneration in the cnidarian Hydractinia. The questions we are interested in are related to how cells make decisions in these contexts. Techniques we use in the lab include random-integration and CRISPR-Cas9 mediated transgenesis/mutagenesis, flow cytometry, cell and tissue transplantation, gene expression analysis, and confocal microscopy for fixed tissues and live imaging experiments.
Cnidarians (sea anemones, corals, and jellies) are emerging model organisms in developmental biology and evolution. Hydractinia symbiolongicarpus, our lab animal, is one of only a few established cnidarian models. The animal grows well in the lab, reproduces sexually every day, and is highly regenerative. A high quality, PacBio based genome sequence is available together with numerous tissue-specific transcriptomes. Hydractinia is small, translucent, and sessile in most stages of its life cycle. This enables in vivo experiments that are very difficult to perform on other animals.
One postdoc will work on characterizing the transcriptomes of all Hydractinia cell lineages at single-cell resolution. The work will also include functional studies on key lineage regulators. The second postdoc will study the transcriptional changes and chromatin landscape that underlie a novel type of regeneration involving natural reprogramming of somatic cells. The positions are funded by NSF and Wellcome, respectively, and are available for three years each.
Candidates must have a PhD in developmental biology, cell biology, or related area. A strong background in molecular biology, experience in working with an animal model, or bioinformatics would be advantageous.
To apply, send a cover letter articulating your interest in one of the projects, your CV, and contact info for at least two references, ideally as a single PDF, to Prof. Uri Frank <uri.frank@nuigalway.ie>. Informal enquiries are welcome.
Stem cells are typically defined by their ability to self-renew and differentiate. These activities are tightly controlled by both intrinsic cues and extrinsic cues from the microenvironment, known as the SC niche. This niche consists of multiple components, among which blood vessels (BVs) are critical as they not only supply oxygen and nutrients to the SCs but also provide molecular signals. BVs form a perivascular-niche for many adult SCs including neural, mesenchymal and hematopoietic SCs. A molecular connection between SCs and vasculature contributes to tissue homeostasis and repair. However, it remains unclear whether this connection also exists in epithelial stem cells, and it’s also unknown whether SCs can conversely promote remodeling of their own environment for proper tissue homeostasis.
The Tumbar lab at Cornell University uses the mouse hair follicle as a model system to study SCs. Hair follicles (HFs) are characterized by a cyclic destruction and reconstruction, which consists of three morphologically distinct and synchronous phases (Figure 1) : 1) growth and proliferation which results into the formation of a new hair shaft known as anagen; 2) apoptosis driven regression, or catagen; 3) and the resting phase, or telogen. In these three phases, SCs exhibits distinct behaviors such as proliferation, migration, or quiescence.
Figure 1: Stem cell behavior during hair cycle in adult skin. Hair cycle is divided into morphologically distinct and synchronous phases: 1) growth and proliferation; 2) apoptosis driven regression; 3) and the resting phase. At the end of quiescence phase stem cells (shown as blue circles) migrate out of their niche (shown as purple crescent) and in response to the activation signals and they change their gene expression, these are called early progenitor cells (shown as red circles). Transcription factor Runx1 is highly expressed in these cells. During the growth or proliferation phase stem cells undergo self-renewal to fill the vacant space and they differentiate and proliferate to make hair shaft. Towards the end of the proliferation differentiated cells undergo apoptosis and stem cells return to quiescence.
A decade ago, when Tudorita (Doina) was a postdoc with Elaine Fuchs studied the transcriptional profile of the hair follicle stem cells (HFSCs) by purifying label-retaining cells in the lineage tracing experiment (Tumbar et al., 2004). Interestingly, many differentially expressed genes in this population encode secreted molecules, suggesting that in addition to receiving signals from the niche, HFSCs may also modulate the niche (Fuchs et al., 2004). More recently, the Tumbar lab has identified runt-related transcription factor 1 (Runx1) as a HFSC regulator. Epithelial Runx1 knockout mice have significant delay in hair growth (Hoi et al., 2010; Osorio et al., 2008). Runx1 is highly expressed in activated SCs or early progenitor cells and its expression is lost when these cells are proliferating (Figure 1). Microarray analysis further revealed gene expression changes in response to altered Runx1 level in the epithelium (Lee et al., 2014).
Our journey started when Prachi, Post-doc in the lab, got intrigued by the microarray data: the gene changes include secreted molecules whose functions are implicated in vascular remodeling, indicating a cross talk between vasculature and stem cell activation. We started to look for blood vessel remodeling in response to varying levels of Runx1 using mutants: Runx1 epithelial knockout mice (Runx1 EpiKO) and Runx1 epithelial transgenic overexpression mice (Runx1 EpiTG). Blood vessel connection in the skin has mainly been studied for oxygen and nutrition, and its molecular signaling aspect has been addressed in depth. Quick experiments of immunostaining for CD31, an endothelial marker, showed visual differences that were exciting and prompted us to begin our investigation on the connection between HFSC and the vascular niche. We were interested in testing whether the vasculature signals for HFSC activation and/or if HFSCs themselves send signals to remodel vasculature during normal hair homeostasis.
Our basic idea was to perturb one compartment, either epithelium or endothelium, and see what would happen to the other compartment. Taking lead from our earlier observation that BVs are remodeled in response to varying levels of Runx1, undergrad student Catherine focused on quantifying the direct contact between vasculature and different regions of the hair follicle. By subsequent quantification, we found that Runx1 mutants showed distinct patterns of vasculature contact. However, these patterns were at first confusing, and we were not sure how to interpret them.
We then began our analysis by quantifying the area covered by CD31 immunostained vasculature under the hair germ and above the muscle. The quantification process was not easy. First, the selection of CD31 immunostained vasculature needs to be manual, because software such as ImageJ or ilastik are not yet as smart as humans in accurately picking out the vasculature. Second, some of the images were not good quality – for example, some stainings appeared hazy, or in some slides the hypodermis was washed away. Those images could be used for contact quantification, but would not be good for area quantification. Therefore, a lot of stainings had to be repeated. Another problem we encountered the different way different researchers quantified their data. The first part of quantification was done by Flora Eun, a brilliant undergrad who soon graduated and left the lab, and the second part of the job was then picked up by graduate student Nina Li. However, Nina was able to be more precise than Flora in selecting the vasculature, so it did not make sense to combine their data as it would create large error within the dataset. To avoid this problem, Nina then quantified the entire dataset, and compared her statistical result with Flora’s preliminary result. Since both statistical results show the same trend that Runx1 EpiKO has significantly more vasculature than control mice, we felt comfortable to further pursue the study.
We then noticed that the thickness of the hypodermis was different in each mouse, but were able to confirm that this thickness did not relate to the amount of vasculature. Since we reasoned that Runx1 might modulate the vascular niche via secreted molecules, the vasculature in close vicinity to the hair germ should be the most affected population, and they should also be the most important if they can send signals back to the hair germ. We used a more stringent method to quantify vascular differences: we drew a thin stripe under the hair germ, and only quantified the vasculature in this selected region. As expected, Runx1 EpiKO had significantly higher amount of vasculature than control mice, suggesting that HFSCs may actively modulate their own vascular niche via Runx1 expression.
In the reverse signaling, from endothelium to epithelium, we did an initial screen before pursuing a particular mutant. From Dr. Anne Eichmann, we acquired two vasculature-related mutants: Cdh5-CreERT2 mediated endothelial knockout mice for Neuropilin 1 (Nrp1) and activin-receptor like kinase 1 (Alk1) genes (Nrp1 EndKO and Alk1 EndKO). Both genes are critical for endothelial cell homeostasis, and perturbation of either one can lead to serious vasculature related diseases. Since we wanted to know whether perturbation of the endothelium affects hair follicle homeostasis, the idea was to perturb BV during quiescence stage before beginning of next stages of HFSCs proliferation and differentiation. When Prachi started optimizing time points and dosage for knockouts she wasn’t sure that if Alk1 conditional knockout out would be viable and show phenotypes, as previously reported knockout pups could only survive 48 hours after birth. Another question was whether adult vasculature patterns can be remodeled by introducing genetic mutations.
We first checked the hair cycle progression on a few mutant mice after knockout induction at PD17, and sacrificed around PD35. While Nrp1 EndKO did not exhibit obvious epithelial phenotype, we were excited to find Alk1 EndKO showed a hair cycle delay phenotype. Alk1 EndKO in the quiescence stage of hair cycle seemed to modulate vascular remodeling and result in delayed progression of stem cell activation. Therefore, we decided to make further investigation on Alk1 EndKO. To confirm the hair cycle delay phenotype, we checked more mice and at various time points between PD22 to PD35. To distinguish between quiescence and early proliferation, since overall morphology does not differ much at these stages, we did both immunofluorescence staining for Ki67, a proliferation marker, and H&E staining to carefully determine the hair cycle stages for our samples. After checking more than 20 mice, we concluded that Alk1 EndKO mice have delayed proliferation marked by the lack of Ki67 staining. We also confirmed that observed delay was not due to less number of early progenitor cells in the mutant mice but rather as a result of delay in HFSC proliferation.
We then wondered what changes in the skin vasculature could lead to the Alk1 EndKO hair cycle delay phenotype. Again, we did immunofluorescence staining for CD31 and we found that there are more CD31+ vasculature in vicinity of the hair germ. The similarity between Alk1 EndKO and Runx1 EpiKO led us to wonder what is the importance of the vasculature near the hair germ. To answer this question, we first investigated the nature of this vasculature during hair homeostasis using wildtype mice (Figure 2).
Figure 2: Skin vasculatures during the hair cycle. The hair cycle consists of three stages. In telogen, both bulge cells and primed stem cells (hair follicle stem cells, HFSCs) in the hair germ remain quiescent. When the hair follicle receives signals to enter anagen, bulge cells proliferate and self-renew, and HFSCs in hair germ proliferate and differentiate to give rise to multiple lineages. In catagen, the regression stage, differentiated lineages generated in anagen undergo apoptosis. Eventually, the hair follicle enters telogen again. During the hair cycle, skin vasculatures (shown as red cables) also change in parallel to the hair follicles. From late catagen to telogen, a horizontal plexus under hair germ (HPuHG) assembles. In anagen, endothelial cells proliferate and disperse. In catagen, endothelial cells undergo apoptosis and deposit to form the HPuHG.
We checked the vasculature arrangement at different stages of the hair cycle. Interestingly, we found that at quiescence stage, a horizontal vascular plexus is formed under the hair germ, which we named it “Horizontal Plexus under Hair Germ (HPuHG)”. This vascular plexus quickly disperses as the hair follicle starts proliferating. Then after apoptosis, the skin vasculature deposits to form the HPuHG again. Since we looked at the quiescence stage in Runx1 EpiKO and in Alk1 EndKO, a stage when HPuHG exists, we believed that the vascular phenotype in the two mutants is in fact an increase in the HPuHG vasculature. These observations strengthened our idea that BVs have distinct role in molecular signaling in addition to nutrition and oxygen, contrary to our belief that increased vasculature causes delay in hair growth. Moreover, distinct pattern of BVs during hair growth supports the idea that BVs are important part of HFSCs niche.
To understand whether molecular signal derived from endothelial cells could cause the HFSC activation defect, we checked if a well-established HFSC quiescence factor BMP4 is also expressed in skin endothelial cells. Previous study indicated endothelial cells express BMP4 in the lung (Frank David et al., 2005). We did immunofluorescence staining for CD31 and BMP4, and found colocalization between the two signals, indicating skin endothelial cells are a source of BMP4. While the expression level of BMP4 in endothelial cells and the average BMP4 intensity in interfollicular epidermis are not altered by the mutation of Alk1 or Runx1, the average intensity of BMP4 in the selected HPuHG region is higher in mutants than in control mice. Western blots also showed elevated BMP4 level in Alk1 EndKO. Though more solid evidence such as RNA-Seq of hair follicle cells from the mutant skin is needed to conclude that endothelial BMP4 is the cause of the HFSC activation delay, our data showed a correlation between BMP4 level, amount of vasculature, and HFSC activation defect. Specifically, our data suggested that excessive HPuHG vasculature may lead to excessive BMP4 in the vicinity of the hair germ, thus delaying the HFSC activation.
There have been previous investigations into the association between the skin vasculature and the hair follicle, but these have focused mainly on hair growth related to vasculature and failed to provide molecular evidence on how the vasculature regulates hair homeostasis (Ellis and Moretti, 1959; Mecklenburg et al., 2000). Our current research focuses on the quiescence stage that has been overlooked for many years, and how cross talk between HFSC and their environment influences their activity. Our data support a model where HFSCs are capable of sending signals to their vascular niche, and the vascular niche can reciprocally control HFSC activation (model summarized in Figure 3).
Figure 3: Hypothetical model of the cross-talk between HFSCs and their vascular niche. Our work suggested that a cross-talk exists between HFSCs and their vascular niche. In our hypothetical model, HPuHG (shown as red cables) starts to form from late catagen to telogen. Endothelial cells near the HPuHG secrete quiescence factors such as BMP4 to maintain the quiescence of HFSCs. In the reverse signaling, HFSCs in the hair germ (shown in green) use Runx1 as a master regulator to modulate their vascular niche. Specifically, many secreted molecules encoded by vasculature-related genes are downstream of Runx1. In response to Runx1 expression, HFSCs remodel the vasculature around them. Proper dispersal of the vasculature leads to the activation of HFSCs.
To our knowledge, this is the first publication suggesting a niche modulation role of HFSCs, and also a first documentation of the non-cell autonomous role of Runx1. Though more molecular details underlying the cross-talk between HFSCs and their vascular niche are to be elucidated, our research has opened a new window for future investigation. The Alk1 mutation has been studied in hereditary hemorrhagic telangiectasia, a disease marked by the fusion between arteries and veins. Our current study may further the understanding of its prevalence in skin, and may also provide new thoughts for future therapy. It also has implications for tissue regeneration and in clinical settings where targeting cancer SC niche through anti-angiogenic targets can be promising therapeutics.
Mecklenburg, L., Tobin, D.J., Müller-Röver, S., Handjiski, B., Wendt, G., Peters, E.M.J., Pohl, S., Moll, I., and Paus, R. (2000). Active Hair Growth (Anagen) is Associated with Angiogenesis. Journal of Investigative Dermatology 114, 909-916.
POST-DOC POSITION IN DEVELOPMENTAL CELL BIOLOGY AND PHYSIOLOGY (3 years in Nice, IBV, France).
Position available (starting early 2020) to functionally characterize the role of Hedgehog in the inter-cellular and inter-organ communication in Drosophila.
Hedgehog proteinsare known key signaling mediators that govern tissue patterning and homeostasis during both development and adult life. The laboratory is interested in how Hedgehog proteins traffic in the producing tissue and exert their function in the receiving tissue, both in a paracrine and hormonal manner.
We have shown that the Endosomal Sorting Complex Required forTransport (ESCRT) promotes Hedgehog proteins loading on exo-vesicles to exert their effect at long distances. We also have shown recently that circulating Hedgehog has a protective role and have identified targets of Hedgehog signaling in glial cells involved in this process. This newly identified role for Hedgehog is important to provide protection during the ageing process. The post-doctoral project aims to gain further insight into the trafficking, vesicular secretion and the extracellular spread of Hedgehog proteins, both at the intercellular and inter-organ level, using cell biology and genetic technics. Invivoimaging and single molecule tracking (in collaboration with computational science lab) has also been developped on our tissue models and will be further usedto investigatethe dynamics of Hedgehog release and spreading.
Interested candidates should have strong knowledge of, and experience in fly genetics, cell biology and optic microscopy (confocal/spinning disc). The position is funded for 3 years in duration. Candidates must have a Ph.D. degree, and can be nationals of any country.
Selectedreferences: Ayers et al., Dev.Cell2010 vol18, 605–620; Briscoe and Thérond, NatRevMolCellBiol.Vol. 14, 2013; Matusek et al., Nature2014 Dec 4;516(7529): 99-103; D’Angelo et al., Dev.Cell2015 Feb. 9 ; 32, 290-303.
Candidates should send a Curriculum Vitae and a list of three referees to:
Welcome to our monthly trawl for developmental biology (and related) preprints.
This month features a series of preprints on stem cell mechanics and tools to help you make organoids, some nectins and some nestins, plenty of auxin in our plant section, and some phantom crustaceans and macabre French genomics in our ‘Why Not’ section.
They were hosted on bioRxiv, PeerJ, andarXiv. Let us know if we missed anything, and use these links to get to the section you want:
Plexin-B2 is a key regulator of cell mechanics during multicellular organization
Chrystian Junqueira Alves, Rafael Dariolli, Theodore Hannah, Robert J. Wiener, Nicolas Daviaud, Rut Tejero, G. Luca Gusella, Nadejda M. Tsankova, Rodrigo Alves Dias, José Paulo R. Furtado de Mendonça, Evren U. Azeloglu, Roland H. Friedel, Hongyan Zou
Transgene-mediated skeletal phenotypic variation in zebrafish
Charles B. Kimmel, Alexander L. Wind, Whitney Oliva, Samuel D. Ahlquist, Charline Walker, John Dowd, Bernardo Blanco-Sánchez, Tom A. Titus, Peter Batzel, John H. Postlethwait, James T. Nichols
Integrating healthcare and research genetic data empowers the discovery of 49 novel developmental disorders
Joanna Kaplanis, Kaitlin E Samocha, Laurens Wiel, Zhancheng Zhang, Kevin Arvai, Ruth Eberhardt, Giuseppe Gallone, Stefan H Lelieveld, Hilary Martin, Jeremy McRae, Patrick Short, Rebecca Torene, Elke de Boer, Petr Danecek, Eugene James Gardner, Ni Huang, Jenny Lord, Inigo Martincorena, Rolph Pfundt, Margot Reijnders, Alison Yeung, Helger Yntema, DDD study, Lisenka Vissers, Jane Juusola, Caroline Wright, Han Brunner, Helen V Firth, David R Fitzpatrick, Jeffrey C Barrett, Matthew E Hurles, Christian Gilissen, Kyle Retterer
Single-cell analysis of human retina identifies evolutionarily conserved and species-specific mechanisms controlling development
Yufeng Lu, Fion Shiau, Wenyang Yi, Suying Lu, Qian Wu, Joel D. Pearson, Alyssa Kallman, Suijuan Zhong, Thanh Hoang, Zhentao Zuo, Fangqi Zhao, Mei Zhang, Nicole Tsai, Yan Zhuo, Sheng He, Jun Zhang, Genevieve L. Stein-O’Brien, Thomas D. Sherman, Xin Duan, Elana J. Fertig, Loyal A. Goff, Donald J. Zack, James T. Handa, Tian Xue, Rod Bremner, Seth Blackshaw, Xiaoqun Wang, Brian S. Clark
Predicting cellular position in the Drosophila embryo from Single-Cell Transcriptomics data
Jovan Tanevski, Thin Nguyen, Buu Truong, Nikos Karaiskos, Mehmet Eren Ahsen, Xinyu Zhang, Chang Shu, Ke Xu, Xiaoyu Liang, Ying Hu, Hoang V.V. Pham, Li Xiaomei, Thuc D. Le, Adi L. Tarca, Gaurav Bhatti, Roberto Romero, Nestoras Karathanasis, Phillipe Loher, Yang Chen, Zhengqing Ouyang, Disheng Mao, Yuping Zhang, Maryam Zand, Jianhua Ruan, Christoph Hafemeister, Peng Qiu, Duc Tran, Tin Nguyen, Attila Gabor, Thomas Yu, Enrico Glaab, Roland Krause, Peter Banda, DREAM SCTC Consortium, Gustavo Stolovitzky, Nikolaus Rajewsky, Julio Saez-Rodriguez, Pablo Meyer
Single-Cell RNA-Seq Reveals Endocardial Defect in Hypoplastic Left Heart Syndrome
Yifei Miao, Lei Tian, Marcy Martin, Sharon L. Paige, Francisco X. Galdos, Jibiao Li, Alyssa Guttman, Yuning Wei, Jan-Renier Moonen, Hao Zhang, Ning Ma, Bing Zhang, Paul Grossfeld, Seema Mital, David Chitayat, Joseph C. Wu, Marlene Rabinovitch, Timothy J. Nelson, Shuyi Nie, Sean M. Wu, Mingxia Gu
Parkinson’s disease phenotypes in patient specific brain organoids are improved by HP-β-CD treatment
Kyriaki Barmpa, Isabel Rosety, Lisa M. Smits, Jonathan Arias-Fuenzalida, Jonas Walter, Gemma Gomez-Giro, Anna S Monzel, Xiaobing Qing, Gerald Cruciani, Ibrahim Boussaad, Christian Jaeger, Aleksandar Rakovic, Emanuel Berger, Paul Antony, Christine Klein, Rejko Krüger, Philip Seibler, Javier Jarazo, Jens C. Schwamborn, Silvia Bolognin
Atypical neurogenesis in induced pluripotent stem cell (iPSC) from autistic individuals
Dwaipayan Adhya, Vivek Swarup, Roland Nagy, Lucia Dutan Polit, Carole Shum, Kamila Jozwik, Paulina Nowosiad, Irene Lee, David Skuse, Eva Loth, Deirdre Howley, Frances A Flinter, Grainne McAlonan, Maria Andreina Mendez, Jamie Horder, Declan Murphy, Daniel H. Geschwind, Jack Price, Jason Carroll, Deepak P. Srivastava, Simon Baron-Cohen
Plants with self-sustained luminescence
Tatiana Mitiouchkina, Alexander S. Mishin, Louisa Gonzalez Somermeyer, Nadezhda M. Markina, Tatiana V. Chepurnyh, Elena B. Guglya, Tatiana A. Karataeva, Kseniia A. Palkina, Ekaterina S. Shakhova, Liliia I. Fakhranurova, Sofia V. Chekova, Aleksandra S. Tsarkova, Yaroslav V. Golubev, Vadim V. Negrebetsky, Sergey A. Dolgushin, Pavel V. Shalaev, Olesya A. Melnik, Victoria O. Shipunova, Sergey M. Deyev, Andrey I. Bubyrev, Alexander S. Pushin, Vladimir V. Choob, Sergey V. Dolgov, Fyodor A. Kondrashov, Ilia V. Yampolsky, Karen S. Sarkisyan
A proximity biotinylation map of a human cell
Christopher D. Go, James D.R. Knight, Archita Rajasekharan, Bhavisha Rathod, Geoffrey G. Hesketh, Kento T. Abe, Ji-Young Youn, Payman Samavarchi-Tehrani, Hui Zhang, Lucie Y. Zhu, Evelyn Popiel, Jean-Philippe Lambert, Étienne Coyaud, Sally W.T. Cheung, Dushyandi Rajendran, Cassandra J. Wong, Hana Antonicka, Laurence Pelletier, Brian Raught, Alexander F. Palazzo, Eric A. Shoubridge, Anne-Claude Gingras
CRISPR-Cas12a-assisted PCR tagging of mammalian genes
Julia Fueller, Konrad Herbst, Matthias Meurer, Krisztina Gubicza, Bahtiyar Kurtulmus, Julia D. Knopf, Daniel Kirrmaier, Benjamin C. Buchmuller, Gislene Pereira, Marius K. Lemberg, Michael Knop
3D super-resolution deep-tissue imaging in living mice
Mary Grace M. Velasco, Mengyang Zhang, Jacopo Antonello, Peng Yuan, Edward S. Allgeyer, Dennis May, Ons M’Saad, Phylicia Kidd, Andrew E. S. Barentine, Valentina Greco, Jaime Grutzendler, Martin J. Booth, Joerg Bewersdorf
Insights from a survey-based analysis of the academic job market
Jason D. Fernandes, Sarvenaz Sarabipour, Christopher T. Smith, Natalie M. Niemi, Nafisa M. Jadavji, Ariangela J. Kozik, Alex S. Holehouse, Vikas Pejaver, Orsolya Symmons, Alexandre W. Bisson Filho, Amanda Haage
Community Standards for Open Cell Migration Data
Alejandra N. Gonzalez-Beltran, Paola Masuzzo, Christophe Ampe, Gert-Jan Bakker, Sébastien Besson, Robert H. Eibl, Peter Friedl, Matthias Gunzer, Mark Kittisopikul, Sylvia E. Le Dévédec, Simone Leo, Josh Moore, Yael Paran, Jaime Prilusky, Philippe Rocca-Serra, Philippe Roudot, Marc Schuster, Gwendolien Sergeant, Staffan Strömblad, Jason R. Swedlow, Merijn van Erp, Marleen Van Troys, Assaf Zaritsky, Susanna-Assunta Sansone, Lennart Martens
Wikidata as a FAIR knowledge graph for the life sciences
Andra Waagmeester, Gregory Stupp, Sebastian Burgstaller-Muehlbacher, Benjamin M. Good, Malachi Griffith, Obi Griffith, Kristina Hanspers, Henning Hermjakob, Kevin Hybiske, Sarah M. Keating, Magnus Manske, Michael Mayers, Elvira Mitraka, Alexander R. Pico, Timothy Putman, Anders Riutta, Núria Queralt-Rosinach, Lynn M. Schriml, Denise Slenter, Ginger Tsueng, Roger Tu, Egon Willighagen, Chunlei Wu, Andrew I. Su
Guidelines for reporting single-cell RNA-Seq experiments
Anja Füllgrabe, Nancy George, Matthew Green, Parisa Nejad, Bruce Aronow, Laura Clarke, Silvie Korena Fexova, Clay Fischer, Mallory Ann Freeberg, Laura Huerta, Norman Morrison, Richard H. Scheuermann, Deanne Taylor, Nicole Vasilevsky, Nils Gehlenborg, John Marioni, Sarah Teichmann, Alvis Brazma, Irene Papatheodorou
A team of neuroscientists led by Professor Christiana Ruhrberg (UCL, UK) and Professor Anna Cariboni (University of Milan, Italy) have found two molecules that work together to help set up the sense of smell and pave the way to puberty in mice. These findings, reported in the journal Development, may help our understanding of why patients with the inherited condition Kallmann syndrome cannot smell properly and cannot start puberty without hormone treatment.
Aficionados of 1990s jazz and fans of David Lynch’s Twin Peaks might remember the distinctive contralto vocals of “Little” Jimmy Scott. Jimmy’s naturally high singing voice was caused by a rare genetic disease, known as Kallmann syndrome, which affects about 1 in 30,000 males and 1 in 120,000 females.
Kallmann syndrome is caused by the lack of a hormone that stimulates the brain to produce signals needed to reach sexual maturity. As a result, people with the condition don’t go through puberty and instead retain a child-like stature, no sex drive and underdeveloped genitals. Currently, the most common treatment is hormone-replacement therapy to bypass the brain and kick-start puberty. Unlike similar reproductive conditions, Kallmann syndrome patients also have no sense of smell – a tell-tale sign of this particular disorder.
Now, research has identified two molecules, called PLXNA1 and PLXNA3, that might be linked to the condition. Scientists have found that both molecules are present in nerves that extend from the nose into the brain of developing mice. These nerves transmit signals essential for the sense of smell and also guide hormone-secreting nerve cells from their place of origin in the nose to their destination in the brain, where they regulate the onset of puberty. The study has revealed that both types of nerve are not wired properly when PLXNA1 and PLXNA3 are absent in developing mice. Consequently, the brain regions that process smells are poorly formed and the brain also lacks the puberty-promoting nerve cells – the same symptoms shown by Kallmann syndrome patients.
“By studying the mouse as a model organism, we have identified a pair of genes that can cause an inherited condition with symptoms similar to human Kallmann syndrome. This is an important finding, because the nerves that convey our sense of smell and that guide the puberty-inducing nerve cells arise in a very similar way during the development of mice and humans whilst they are still in the womb,” explained Professor Christiana Ruhrberg, who led the UK team.
This research gives hope to patients with an unknown cause of Kallmann syndrome by testing for defects in the PLXNA3 gene together with PLXNA1, which has been previously implicated. The lead author from the University of Milan, Professor Anna Cariboni added, “Although Kallmann syndrome can be treated with hormone injections if diagnosed early, knowing the underlying genetic causes can make a huge difference to speed up diagnosis and give treatment to the right patients at an earlier time.”
The image contains a circular structure within the developing nose that gives rise to puberty-inducing nerve cells, shown in green. Some of these nerve cells leave their birth place to travel in small clumps along nerve cables. These cells and the cables also have the PLXNA3 molecule, which is coloured red, but appears yellow, because of the red and green overlap. All cells are also highlighted in blue. CREDIT: Roberto Oleari, University of Milan.
Pre-trained Models for Developmental Biology Authors: Bradly Alicea, Richard Gordon, Abraham Kohrmann, Jesse Parent, Vinay Varma
Our virtual discussion group (DevoWormML) has been exploring a number of topics related to the use of pre-trained models in machine learning (specifically deep learning). Pre-trained models such as GPT-2 [1], pix2pix [2], and OpenPose [3] are used for analyzing many specialized types of data (linguistics, image to image translation, and human body features, respectively) and have a number of potential uses for the analysis of biological data in particular. It may be challenging to find large, rich, and specific datasets for training a more general model. This is often the case in the fields of Bioinformatics or Medical Image analysis. Data acquisition in such fields is often restricted due to the following factors:
* privacy restrictions inhibit public access to personal information, and may impose limits on data use.
* a lack of labels and effective metadata for describing cases, variables, and context.
* missing data points, which require a strategy to normalize and can make the input data useless.
We can use these pre-trained models to extract a general description of classes and features without requiring a prohibitive amount of training data. We estimate that the amount of required training data may be reduced by an order of magnitude. To get this advantage, pre-trained models must be suitable to the type of input data. There are a number of models specialized for language processing and general use, but options are fewer within the unique feature space of developmental biology, in particular. In this post, we will propose that developmental biology requires a specialized pre-trained model.
This vision for a developmental biology-specific pre-trained model would be specialized for image data. Whereas molecular data might be better served with existing models specialized for linguistic- and physics-based models, we seek to address several features of developmental biology that might be underfit using current models:
* cell division and differentiation events.
* features demonstrating the relationship between growth and motion.
* mapping between spatial and temporal context.
Successful application of pre-trained models is contingent to our research problem. Most existing pre-trained models operate on two-dimensional data, while data types such as medical images are three-dimensional. A study by Raghu et.al [4] suggests techniques specified by pre-trained models (such as transfer learning by the ImageNet model) applied to a data set of medical images provides little benefit to performance. In this case, performance can be improved using data augmentation techniques. Data Augmentation, such as adding versions of the images that have undergone transformations such as magnification, translation, rotation, or shearing, can be used to add variability of our data and improve the generalizability of a given model.
One aspect of pre-trained models we would like to keep in mind is that models are not perfect representations of the phenomenology we want to study. Models can be useful, but are often not completely accurate. A model of the embryo, for example, might be based on the mean behavior of the phenomenology. Transitional states [5], far-from-equilibrium behaviors [6], and rare events are not well-suited to such a model. By contrast, a generative model that considers many of these features might generally underfit the mean behavior. We will revisit this distinction in the context of “blobs” and “symbols”, but for now, it appears that models are expected to be both imperfect and incomplete.
The inherent imperfection of models is both good and bad news for our pursuit. On the one hand, specialized models cannot be too specific, lest they overfit some aspects of development but not others. Conversely, highly generalized models assume that there are universal features that transcend all types of systems, from physical to social, and from artificial to natural. One example of this is found in complex network models, widely used to represent everything from proteomes to brains to societies. In their general form, complex network models are not customized for specific problems, relying instead on the node and edge formalism to represent interactions between discrete units. But this also requires that the biological system be represented in a specific way to enforce the general rules of the model. For example, a neural network’s focus on connectivity requires representations of a nervous system to be simplified down to nodes and arcs. As opposed to universality, particularism is an approach that favors the particular features of a given system, and does not require an ill-suited representation of the data. Going back to the complex networks example, there are specialized models such as multi-level networks and hybrid models (dynamical systems and complex networks) that solves the problem of universal assumptions.
Another aspect of pre-trained models is in balancing the amount of training data needed to produce an improvement in performance. How much training data can we save by applying a pre-trained model to our data set? We can reformulate this question more specifically to match our specific phenomenon and research interests. To put this in concrete terms, let us consider a hypothetical set of biological images. These images can represent discrete points in developmental time, or a range of biological diversity. Now let us suppose a developmental phenotype for which we want to extract multiple features. What features might be of interest, and are those features immediately obvious?
In the DevoWorm group (where we mostly deal with embryogenetic data), we have approached this in two ways. The first is to model the embryo as a mass of cells, so that the major features of interest are the shape, size, and position of cells in an expanding and shifting whole. Last summer, we worked on applying deep learning to
While these models were effective for discovering discrete structural units (cells, filaments), they were not as effective at directly modeling movement, currents, or transformational processes. The second way we have approached this is to model the process of cell division and differentiation as a spatial and discrete temporal process. This includes the application of representational models such as game theory [7] and cellular automata [8]. This allows us to identify more subtle features that are not directly observable in the phenotype, but are less useful for predicting specific events or defining a distinct feature space.
Our model must be capable of modeling multiple structural features concurrently, but also sensitive to scenarios where single sets of attributes might yield more information. Ideally, we desire a training dataset that perfectly balances “biologically-typical” motion and transformations with clearly masked shapes representing cells and other phenotypic structures. Generally speaking, the greater degree of natural variation in the training dataset, the more robust the pre-trained model will turn out to be. More robust models will generally be easier to use during the testing phase, and result in a reduction in the need for subsequent training.
Finally, specialized pre-trained models bring up the issue of how to balance rival strategies for analyzing complex processes and data features. Conventional artificial intelligence techniques have relied on a representation which relies on the manipulation of symbols or a symbolic layer that results from the transformation of raw data to a mental framework. By contrast, modern machine learning methods rely on data to build a series of relationships that inform a classificatory system. While a combination of these two strategies might seem obvious, it is by no means a simple matter of implementation [9]. The notion of “blobs” (data) versus “symbols” (representations) draws on the current debate related to data-intensive representations versus formal (innate) representations [10-12], which demonstrates the timeliness of our efforts. Balancing these competing strategies in a pre-trained model allows us to more easily bring expert knowledge or complementary data (e.g. gene expression data in an analysis of embryonic phenotypes) to bear.
We will be exploring the details of pre-trained models in future discussions and meetings of the DevoWormML group. Please feel free to join us on Wednesdays at 1pm UTC at https://tiny.cc/DevoWorm or find us on Github (https://github.com/devoworm/DW-ML) if you are interested in discussing this further. You can also view our previous discussions on the DevoWorm YouTube channel, DevoWormML playlist (https://bit.ly/2Ni7Fs2).
[2] Isola, P., Zhu, J-Y., Zhou, T., Efros, A.A. (2017). Image-to-Image Translation with Conditional Adversarial Nets. Proceedings of Conference on Computer Vision and Pattern Recognition (CVPR).
[3] Cao, Z., Hidalgo, G., Simon, T., Wei, S-E., and Sheikh, Y. (2018). OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields. arXiv, 1812.08008.
[4] Raghu, M., Zhang, C., Kleinberg, J.M., and Bengio, S. (2019). Transfusion: Understanding Transfer Learning for Medical Imaging. arXiv, 1902.07208.
[5] Antolovic, V., Lenn, T., Miermont, A., Chubb, J.R. (2019). Transition state dynamics during a stochastic fate choice. Development, 146, dev173740. doi:10.1242/dev.173740.
[6] Goldenfeld, N. and Woese, C. (2011). Life is Physics: Evolution as a Collective Phenomenon Far From Equilibrium. Annual Review of Condensed Matter Physics, 2, 375-399. doi:10.1146/annurev-conmatphys-062910-140509.
[7] Stone, R., Portegys, T., Mikhailovsky, G., and Alicea, B. (2018). Origins of the Embryo: Self-organization through cybernetic regulation. Biosystems, 173, 73-82. doi:10.1016/j.biosystems.2018.08.005.
[8] Portegys, T., Pascualy, G., Gordon, R., McGrew, S., and Alicea, B. (2016). Morphozoic: cellular automata with nested neighborhoods as a metamorphic representation of morphogenesis. In “Multi-Agent Based Simulations Applied to Biological and Environmental Systems“. Chapter 3 in “Multi-Agent-Based Simulations Applied to Biological and Environmental Systems”, IGI Global.
[9] Garnelo, M. and Shanahan, M. (2019). Reconciling deep learning with symbolic artificial intelligence: representing objects and relations. Current Opinion in Behavioral Sciences, 29, 17–23.
[10] Zador, A.M. (2019). A critique of pure learning and what artificial neural networks can learn from animal brains. Nature Communications, 10, 3770.
(1 votes)
Loading...
(No Ratings Yet)
In the latest 
