|Conference and seminar notes|| Main project page|
Previous entry Next entry
Notes for Collège de France seminar series – invited speaker Marianne Bronner-Fraser from CalTech
Marianne speaks a perfectly serviceable French but is giving these talks in English.
May 5th 2010
Good introduction by Alain Prochaintz touching on some of her contributions, in particular gene regulatory networks using the lamprey and Xenopus.
Intro slide with neural crest contribution to pigmentation and faces of different vertebrates (colored, furry etc.)
1/3 of congenital birth defects are neural crest-related. Dysregulated is a word. Prone to transformation – tumors such as glioblastoma, neuroblastoma and melanoma. Why cells do what they do in the embryo to understand what can go wrong later on.
Evolution of neural crest – simplified deuterostome tree; crest at separation between agnathia and urochordates.
Evocation of EMT using Nat Neurosci review illustration of plate to NCC.
Orientation relative to HH10 chicken embryo – after NT closure shape change and NCC migrate out, placed lateral to Slug/SNAI2 – premigratory and postmigratory cells "expanding as a cobra's hood does" all through the head.
Electroporation of neural tube with GFP and then made a beautiful movie of them migrating out of the rhombencepalon – it's gorgeous. Can follow the cells individually. Glance over at Sophie Creuzet who gave me a thumbs up. She's my electroporation reference.
Neural crest derivatives and discussion of stem cell potential at the time they are in and just after leaving the neural tube.
Induction NNE/NE by growth factors (N, Wnt, BMP, Fgf), turning on broad-definition transcription factors at wide borders (Msx, Pax, Zic, Dlx3/5), in turn regulating other "neural crest specifier" transcription factors (Slug, FoxD3, cMyc, Twist, AP2, Id, Sox9, Sox10) leading to migration away from NT. Downstream targets eg. differentiation factors and adhesion (Col2a, cRet, Trp2, cadherins) to get migratory cells. "Lots of embryological data to get the description of each one of these steps" – to get a global overview, proposed gene regulatory network over time. Bring in specifics. Would like to draw like an electronic circuit to have all inputs and possible outcomes to get feed-forward movement.
Presented first the broad lines, then the individual examples at each level 1-4. Schematic with Epidermis over column on left, Neural plate border in middle (with list of levels 1-4) and Neural Plate on the right.
Co-explants of NNE and NE not at neural fold = NCC. Wnt6 is in ectoderm. Blockage at fold blocked NCC formation. Add Wnt6 to NE get NCC.
Pax7 – cf expression pattern at HH8+ in the neural folds, all the way around presumptive neural plate at gastrula level (node). Fluorescent ISH in the chick embryo – cf. [Martin Garcia-Castro http://www.ncbi.nlm.nih.gov/pubmed/16688176] - See stage 4+ posterior border of the radial neural plate ISH of Pax7 in crescent. "This is just a marker so it can't tell you anything about what those cells can do" – so "specification assay" – take cell out of context, without added exogenous factors and ask if they go ahead and become what they would have done. Made very fine explants of different regions across st4+ embryos (the "keyhole" stage") If cells specified so early, why does it take so long to execute their migratory program?
Is Pax7 necessary for the specification? Knockdown with antisense morpholinos of Pax7. Electroporated the morpholino into area known by fate map to give rise to neural fold. Loss of Slug expression (also Sox9 and Sox10, but not FoxD3) in treated neural fold. Really NICE whole-mount fluorescent microscopy of the GFP in the later neural fold, before the ISH.
Regulatory regions of chick regions – conserved areas about 1/3 the spread of mammals eg. mouse. Example of cis-regulatory analysis of Sox10. Has inputs into almost all the NCC-derived lineages. The only lineage w/o Sox10 input is cartilage which uses Sox9 instead.
Comparative sequence analysis – Lots of highly conserved regions mostly upstream and sometimes downstream.
Reporter constructs – CMV-IE---RFP generic expression and Putative enhancer-Tkpro-Egfp-pA. Reporter analysis of different fragments "à l'ancien" – enhancer downstream of Sox10 "Sox10E" recapitulates expression of Sox10 in peripheral ganglia, OT, early cranial NCC expression… Dissection of 3.5kb fragment downstream into three fragments, and were able to subdivide the cranial vs vagal/trunk migrating cells according to which subfragment was analyzed.
Bioinformatics to look at activity – of the element from ggcaagagtggcaatttaacctacaactgctgagcttg…. to cacaaaggcccaactgtctaggggaagcaat (presumably [this is published http://www.ncbi.nlm.nih.gov/pubmed/20139305]; lots of TF potential sites – mutated individually and in combination – grad student "Paola" – SoxE (Sox9/10-binding), Ets, and (must mutate both sites:) Myb-binding appeared important to maintenance of Sox10-like expression with this enhancer region. Eliminates reporter expression in cranial neural crest. How do they target electroporation to get the whole head?
Ets1 expression is very local to cranial NCC, and cMyb is early in cranial neural crest and folds. Then used morpholinos against Sox9, lose reporter expression not by mutating site but by knocking down Sox9 expression. DNA pull-downs and gel shifts – all factors do indeed bind this 264-bp fragment: in order, cMyb, SoxE, Ets1, SoxE again and cMyb again. These control the onset of endogenous Sox10 expression in the cranial NCC.
Looked with ISH after knocking down each, then all three with the MOs – triple KO is completely effective, the others were partial (mostly).
Enhancers also useful for tissue-specific knockdown and expression, also for over-expression, as well as cell-sorting NCC populations.
Enhancers of FoxD3 expression upstream of FoxD3 – very different enhancers, but spatiotemporal information still encoded similarly (subdividable). Double fluorescent ISH in toto – I must get this technique. Much less FoxD3 enhancer expn in NCC than Sox10 – time-lapse cinematography during migration shows one early enhancer only in NT and early cells, then another comes on again in some of the Sox10-expressing NCC.
hES into hNCC-like cells. Reporter constructs work in these cells (FoxD3 and Sox10).
MOs are expensive, shRNA construct for tissue-specific gene knockdown. Can be toxic b/c non-endogenous, massive overexpression. Have other ways of shRNA expression purportedly better.
[Laura Gammill's http://www.ncbi.nlm.nih.gov/pubmed/12421712] work: screen for genes involved in NC development, subtractive screen. 900 genes, of which 1/3 expressed in NC by ISH. When a new postdoc comes into lab, look at all these new genes we have – takes about a year to work on a gene. Studied 5 more specifically. Each gene picked from the screen so far has had a NC phenotype. Adh5, Elk3, Ccar1 – all when knocked down, all effects on NC development.
Pablo Strobl wants to work on chromatin remodelers. Wanted to look at jumanji histone demethylase – epigenetics in NC formation. Jumanji demethylates lysine 9, leading to gene expression de-repression.
(This work appears to be unpublished as yet.)
ISH – in neural plate, then neural folds, then v low in migrating cells. JmjD2A-MO perfect knock-down of Sox10 in migrating neural crest cells. True for FoxD3 (migrating cells, not in the neural folds), Snail2/Slug (dimunition, not abolition) but not for Sox2 in the neural tube.
ChIP at HH4-5 or HH10-11. Antibody to H3K9me3, highly methylated near TSS of Sox10 very early on. And Sox10 is indeed off. But at HH10, the lysine is no longer methylated. Amplifies different fragments at -1kb, -0.5kb, +0.5kb, +1kb. The fold enrichment goes from 1-200x! Measures as % input as well. At HH8, Jmj bound upstream to Sox10, not at HH9 anymore. Compare with histone methylation."high Me at HH8, not at HH9" but "high" is 0.2% input to 0.3% input as expressed on y axis. The error bars are pretty reproducible but no idea how those are generated. Still underway, I expect.
JmjD2A removes the Me from the promoter at -500bp even when the neural crest is specified and the TFs around (eg Pax7), at a later stage which allows them to bind finally. JmjD2A represses eg Sox10 at early stages "maintaining stemness".
Pax7 – Sox9/Ets1/cMyb directly to Sox10. JmjD2A removing Me-K9 allows the others to bind.
Not only does MBF acknowledge, but she publishes with some of these people who need their careers to begin, in last-but-one position, which is very good of her:
Martin Basch Paola Betancur Tatjana Sauka-Spengler (movies!!) Pablo Strobl Martin Garcia-Castro (Pax7) Katherine Fishwick (shRNA constructs) Sonja McKeown for FoxD3
NLD: differences in enhancers for cranial/trunk NCC – potential of the cells? Answer – Ets1 is only in head. Regulatory region controlling Ets1 expression. If you cut down the region, there is a repressor keeping Ets expression off in the trunk. Trunk NCC is kept in repressed state relative to the cranial NCC – evolutionarily this would make sense cf. dermal skeleton which was a NCC derivative from truncal neural crest.
2nd question about stemness problem. Sox10-expressing NCC – some of these must remain stem cells as far as behavior. Answer – induced cells in gastrula probably have broader potential and also like neural stem cells. Epigenetic factors. Maybe not a NCC until it expresses Sox10? = NC precursor before that. Perhaps an earlier form of NC precursor which is
JmjD2a also works on FoxD3 and Snai2, not on another TF that I can't remember. Looking also at Msx, Ets-1. Another DNA demethylase seems to play role in NT to NC transition.
Anne-Helene Monsoro-Burq: This particular Jmj is quite NC-specific. Regulation of Jmj perhaps by Wnt6? Not yet looked at it. But must be another earlier Wnt if indeed this is happening at
Luc Pardanaud: TFs regulating Sox9 rather than Sox10? Regulatory region is much harder to dissect. Robin Lovell-Badge working on this. Perhaps has a NC enhancer, not known if teased apart yet.