STEVENS, Beth, Microglia Prune Synapses, Harvard Med, Oct 6, 2016
<p class="MsoNormal" >STEVENS, Beth, Microglia Prune Synapses, Harvard Med, Oct 6, 2016</p>
<p>
<b>full title is how do microglia prune specific synapses in health and disease?...</b></p>
<p>
This is at Stanford in Clark Auditorium at noon (in the SNI series)
this is Beth Stevens, PhD of Boston Children's Hospital and Harvard Med.
About:
The Stevens lab is interested in interactions between the two fundamental cell types of the nervous system, <b>neurons and glia.</b> The laboratory seeks to understand how <b>neuron-glia communication </b>facilitates <b>the formation, elimination and plasticity of synapses</b> — the points of communication between neurons — during both healthy development and disease.
host and intro by David Prince (star emeritus prof)</p>
<p>
(Prince) she used to work on Schwann cells at NIH with Fields then came to</p>
<p>
Ben Barres Lab in 2004 (here.)</p>
<p>
worked with Carla Shatz.</p>
<p>
Beth: big idea: the <b>immune system is intimately involved with synapse removal in the brain</b>.</p>
<p>
<b>C1Q (1st element in complement cascade is involved in synapse pruning</b>.</p>
<p>
mice get epilepsy if too many synapses (in a KO mouse)</p>
<p>
Beth went to Childrens Hosp in 2008 at Harvard Med (Broad Institute).</p>
<p>
and at the FM Kirby Neurobio Center.</p>
<p>
she got a Presidential Award and an early career award.</p>
<p>
this is a mini-glial symposium today (also see my notes on Dwight Bergles 3 PM talk today)
(bob: also see private chats with Beth, Dan Lysko, David Prince, and John Huguenard, and Roula today)</p>
<p>
*
Beth begins.........</p>
<p>
early synapse loss... involved in Alzheimers and in other diseases.</p>
<p>
<b>microglia are involved in normal developmental pruning.</b></p>
<p>
how did this work get started (and then on to unpub work)</p>
<p>
<b>so how do they (the microglia) know WHICH synapses to prune</b>?
(this work is all in mice)</p>
<p>
synaptic pruning in the dev of the brain...</p>
<p>
<b>brain starts off with an excess of synapses (lot of work from Carla Shatz Lab)</b></p>
<p>
see <b>Huttenlocher in 1979</b>.</p>
<p>
<b>necessary for precise brain wiring </b>(must have precise pruning)</p>
<p>
from Schizo to autism... this process is invovled...</p>
<p>
but also diseases of the aged brain...</p>
<p>
in Alzherimers and in FTD= frontotemporal dementia and in Huntingtons.</p>
<p>
so microglia are involved in both neurodev disease and in neurodegen dis of aging.</p>
<p>
<b>big Q: how are CNS synapses eliminated? (it is NOT random)</b></p>
<p>
shows a cartoon... <b>the weak ones get eliminated...</b></p>
<p>
<b>neuronal activity is crucial... sets up a local COMPETITION</b>...</p>
<p>
in her cartoon, the weak (blue ones) get eliminated leaving just the strong red ones.</p>
<p>
MICROGLIA are implicated</p>
<p>
<b>they are 10% of cells in the brain (of total cells)</b></p>
<p>
come from the yolk sac... and take up residence in the brain (all this is in mouse)</p>
<p>
they are the FIRST GLIAL cells in the brain...</p>
<p>
<b>so, what are they surveying??? SYNAPSES themselves</b>...</p>
<p>
great video</p>
<p>
from Jenelle Wallace... get this.</p>
<p>
How do they prune?</p>
<p>
<b>they are resident phagocytes</b>... they eat stuff!</p>
<p>
are they eating them?</p>
<p>
mouse visual system:</p>
<p>
a model for studying synaptic pruning...</p>
<p>
look at lat geniculate...</p>
<p>
like Carla S's work...</p>
<p>
are microglia nibbling off the lat genic inputs... yes!</p>
<p>
see Schafer in Neuron 2012...</p>
<p>
look at LGN</p>
<p>
mcroglia engulf presynaptic inputs during peak eye segregation (also in hippocampus and cortex)</p>
<p>
but do they care WHICH synapses?</p>
<p>
we manip activity with tetrodotoxin and do the reverse with forskolin...</p>
<p>
<b>microglia DO engulf the less active synapses</b>.</p>
<p>
shows that the</p>
<p>
classical complement cascade mediates the CNS synapse elimination... pub in Cell in 2007
she was 1st author.</p>
<p>
<b>complement proteins TAG apoptotic cells and pathogens for rapid elimination...</b></p>
<p>
<b>we showed that C1Q is involved and also CR2</b>...</p>
<p>
<b>microglial engulfment of RGC inputs is dependent on CR3 and C3...</b></p>
<p>
this is ONE of the ways that they prune (but there are also other ways too)</p>
<p>
<b>hyperwired mice are more prone to seizures... </b>hyperconnected is bad...</p>
<p>
we showed that <b>microglia preferentialy engulf less active synapses and this is C1q dependent</b>.</p>
<p>
<b>C1Q is bound preferentially to the less active synapses</b>.</p>
<p>
model 1: is there specific tagging (but how does that occur . is it via a specific receptor)
ie tagged for destruction or</p>
<p>
or is it model 2: <b>specific protection...</b></p>
<p>
maybe it's a <b>combo of protection of the good synapses and attack on the weak ones</b>.</p>
<p>
is that it, she asks.</p>
<p>
are there protective molecules...</p>
<p>
why aren't the microglia just nibbling off the entire dendrite?</p>
<p>
(Irv Weissman is seated next to me... our expert on complement)</p>
<p>
"don't eat me" signals prevent inappropriate pruning.</p>
<p>
cone in on CD47 which directly inhibits phagocytosis...</p>
<p>
microglia expresssion is the
work of Emily Lehrman...in Beth's lab.</p>
<p>
where and when is CD47 expressed...
it is enriched in the dLGN during peak pruning...</p>
<p>
yes the CD47 is expressed a lot during the peak of pruning...</p>
<p>
immunofluor...</p>
<p>
it is colocalized with engulfment... but is it causal?</p>
<p>
does CD47-SIRPalpha do this.</p>
<p>
yes, CD47KO microglia engulf more RGC inputs than WT microglia...</p>
<p>
CD47 loss does enhance pruning.</p>
<p>
we compared ipsi with contralat. in visual system.</p>
<p>
then we went to physiology...</p>
<p>
work of Chinfei Chen...</p>
<p>
make acute slices and stim ex vivo... (in slice)</p>
<p>
in the P60 mouse...</p>
<p>
<b>big reduction in AMPA current and NMDA receptor levels</b></p>
<p>
<b>get an over pruning with loss of CD47</b>.</p>
<p>
yes, indeed, there is an over-pruning with loss of CD47 (in the KO mice)</p>
<p>
they did indeed have fewer retino-genic synapses.</p>
<p>
now let's go back to the microglia</p>
<p>
how about just getting rid of the receptors.</p>
<p>
the CD47-SIRP-alpha...</p>
<p>
it is highly expressed by microglia in the dLGN during the pruning period.</p>
<p>
we showed that SIRPalpha is highly expressed by microglia in the dLGN during the pruning period.</p>
<p>
Emily did this work.</p>
<p>
yes... this is due to an interacditon of microglia and CD47.</p>
<p>
CD47 serves as a "don't eat me" signal that inhibits
SIRP-alpha mediated synaptic phagocytosis by microglia.</p>
<p>
we suggest...</p>
<p>
a ying and yang model...</p>
<p>
removes the presynaptic terminals...</p>
<p>
complement + removal of protection.</p>
<p>
(but what is the receptor)</p>
<p>
unpub work:</p>
<p>
want to see this in real-time... the engulfment of these synapses.</p>
<p>
in disease may see too much or too little pruning...</p>
<p>
eg in schizo...</p>
<p>
hard to tease out cause and effect in schizo...</p>
<p>
is it the balance of protective vs punishment signals to the synapses.</p>
<p>
see Chromosome 6 in schizo...</p>
<p>
work of Aswin Sekar and Steve McCarroll at Hvd Med School at Broad Instit.</p>
<p>
they looked at complement C4.</p>
<p>
two genes: C4A and C4B...</p>
<p>
incredible range of haplotypes for these genes... big diversity in struc variation..</p>
<p>
does this contrib to schizo risk?
NATURE: by Aswin Sekar..</p>
<p>
schizo risk from complex variation of complement component 4.</p>
<p>
alleles of C4 shape schizo risk protection...due to the effect on C4A expression.</p>
<p>
PI is Steven McCarroll.</p>
<p>
hypothesis: excessive pruning contribs to the pathobiology of schizophrenia.</p>
<p>
could over-activ of complement-mediated pruning be contrib to schizo.</p>
<p>
why the onset in adolescence.</p>
<p>
pruning and myelination in late teens ...</p>
<p>
just testing this hypothesis...</p>
<p>
is it aberrant pruning in adolesc.?</p>
<p>
*</p>
<p>
now how about the adult, aged brain...</p>
<p>
AGED BRAINS and neurodegen disease...</p>
<p>
are the above pathways contrib to abnl loss in synapses.</p>
<p>
she focuses on her work on Huntingtons.</p>
<p>
Science in 2016... from her lab</p>
<p>
complement and microglia mediate early synapse loss in Alzheimers mouse models</p>
<p>
1st author is Soyon Hong...</p>
<p>
these mice have way more plaque.</p>
<p>
shows immunfluro of C1q and PSD95 ...</p>
<p>
shows autopsy speciment of Huntingtons...</p>
<p>
affects the cortico-striatal pathway..</p>
<p>
charac by invol movements, dementia and on psychiatric probs.</p>
<p>
degen of cortico-striatal pathway in HD = Huntingtons..</p>
<p>
microglia do preferentially affect synapses...</p>
<p>
but what's the mech of selective vulnerability?</p>
<p>
so.. do aberrant complement and micro-glia mediated pruning contrib to syanpse loss and dysfunction in Huntingtons?</p>
<p>
are synapses lost in the vulnerable regions...</p>
<p>
label Vglut1 and Homer...</p>
<p>
VGlut1 specifically labels cortical-striatal path.</p>
<p>
(not in HC)</p>
<p>
but not in the vGLUT2 syanpses (that are in the cortico-thalamic path)</p>
<p>
contrast control caudate nucleus vs HD2 caudate nucleus (in Huntingtons)</p>
<p>
we DO see striking specific.. region specific upreg of complement but not in dentate gyrus etc.</p>
<p>
complement colocalizes with vulnerable synapses in Huntington mouse model.</p>
<p>
it is selective tagging with complement in vGLUT1 and not in vGLUT2)</p>
<p>
also collab with Richard Faull in New Zealand.. world's bggest Huntinton brain bank.</p>
<p>
do see incr in complement tagging in vGLUT1 positive synapses but not in cerebellum.</p>
<p>
and how early does it happen?
do microglia contrib to synaptic loss in Huntington's</p>
<p>
(they ARE mopping up, but are they contrib early in the disease process)</p>
<p>
shows human brains (immunofluor)</p>
<p>
yes.. incr complement..</p>
<p>
...</p>
<p>
microglia doe engulf neuronal connections in Huntingtons mouse models
in the 3 to 4 month wiindow in the mice... they do have incr engulfment.</p>
<p>
can we block CR3 or complement itself to slow this process down..</p>
<p>
<b>Annexon</b>.. a company founded by Ben Barres and by Arnon Rosenthal..
their product is ANZ005</p>
<p>
the blockage does lead to decrease in C3...</p>
<p>
after one month rx with blocking Abx...</p>
<p>
we do see a signif decr in loss of cortico-striatal synapses...</p>
<p>
(work in progress: can we rescue these pathologies in Huntingtons)</p>
<p>
what triggers this pathway to begin with...</p>
<p>
perhaps its the aberrant firing that is turning this process on..</p>
<p>
but we dont want to affect complements action in other sites. (only at the aberrant sites)</p>
<p>
so, complent-synapse elimination pathways as a novel therapeutic?</p>
<p>
perhaps complement-inhibitors and blocking antibodies
(might be used for therapy)</p>
<p>
might be a common pathway (and might be happening early) in many disease.. Alzheim, glaucoma, FTD e tc.</p>
<p>
many foundations support this work.</p>
<p>
huge lab.</p>
<p>
*</p>
<p>
but how will we make it more specirfic (audience asks)</p>
<p>
A: what is complement binding to? profile microglia with RNA seq .</p>
<p>
Irv W asks... what gets engulfed...(just the preterminal bouton or ... more?</p>
<p>
A.. its the presynaptic terminals
microglia do not engulf the entire arbor...</p>
<p>
implies that this is regional localiz of CD47 (comments Irv)</p>
<p>
she says this is one of many ways that you prune...</p>
<p>
even in our KOs there is still pruning... she comments</p>
<p>
David Prince: how about the astrocytes? are they still thought to be involved
(even though they are now off the front page, he jokes)</p>
<p>
A: she says YES ... astros are involved (I did not mean to de-emphasize them...</p>
<p>
two separte mechanisms... astros have their specificity...</p>
<p>
there is also a motility issue (in microglia)
eg ATP could be one of those cues.</p>
<p>
</p>