Ed Callaway Rabies-based tools for elucidating neural circuits
<p class="MsoNormal" >Nov 06, 2012 07:12 PM
Ed CALLAWAY  video
This is superb work!
Keynote lecture at Neuroinformatics 2011 in Boston, USA.
Ed Callaway, Crick-Jacobs Center for Theoretical and Computational Biology,
Systems Neurobiology Laboratories, Salk Institute for Biological Studies, San Diego, USA.
 
(I just heard Ed speak at Stanford on Nov 5 in Herrin Hall... also see my notes 
on that lecture)
 
first intro:
cud have  best wiring diagram  and it will not tell you how it works;
we need tools to allow direct causation/ correl of circuits...
neurons> microcircuits>modules> areas
what is transformation  and how does it happen...
how do microcircuit mechs transform info  and what gets transmitted to distal areas...
brain slices... isolated preps... do not show connections  distal areas...
Andreas Burkhalter (wash univ st louis),
heroic studies... 
Kevin Martin  used  overlap as clue to connections  (closeness increases prob. of synapse... but can be wrong)
neural connex are cell-type specific
connections are cell-type specific...  the rule rather than the exception....
BUT  sometimes  neurons of the same cell type (even neighbors)  may be connected differently.
take 2 pyrams right next to one another... they are not connected to one another... but
they can share inputs...   but they can serve difft functions...
look at L2/3  pyrams... even right next to one another...  only connected to one another perhaps 10 to20% of time.
we need tools that reveal connectivity and allow direct causal and correl studies linking circuitry to function
in THE SAME ANIMAL  (cuz may differ from animal to animal)
Connectome...
do we we need another ...OME?
purposes of connectome:
to reveal organizing principles (of connex) and
to suggest functional relationships and guide func assays...
but key question is... are benefits commensurate with the costs? (in elucidating the connectome)
this worked out for C elegans
 and for mammalian primary visual areas  eg as in felleman and van essen 91...
eg hierarchy is organizing principle in fell. and van ess. and the
prin. of reciprocal connex 
hierarchy:  if u lesion lower area, it has bigger effect on higher level area than vice versa.
*
The C elegans complete wiring diagram... based on EM  existed for 25 years...
but was much less useful... till just recently
can't stick an electrode in a neuron in c elegans...
 but changed dramatically  by cori Bargman...
but identical in all worms....  so 25 years later it is quite useful... same in every worm...
how abt (EM reconstruction of)  1mm cube of mammalian brain:
 differences from animal to animal and species:
might find connex motifs... but wud not give u functional assays...
 not as valuable  cuz of differences.
if u see  specialization pre and post synapse (active zone eg)... then u know there's a synapse there...
but  axons may bifurcate, cross... hard to follow
RABIES VIRUS...  we use it to id synapse and trace  wires... only spreads thru synapses..
solves our problems (that occur with EM)...  can trace wires and id connections...
cre-recombinase...  express selectively...
cell types are conserved across animals and species... he shows chandelier cell...
same in giraffe, monkey, human...
rules of cell type connectivity are conserved across indiv brains and species...
cell types can be targeted with genetic manips...
single cell connectome...  the complete set of connex for a single cell...
can do rabies system in live animal to see connex of a single cell and do func studies...
he shows brain slice...
a starter cell  (infected with rabies virus)  then the green cells are  rabies infected
(spread only in retrograde direction)...
natural rab. affects axon terminals...  natl rabies would spread all over  (and wud be useless for tracing connex);
need specific infection and need it to go only one synap connec and then stop.
build a better rabies virus
1) want virus to infect spec cell type
2)  want to spread and replicate and
3) only go one step...
rabies virus with glycoprotein gene deleted...
2 kinds of g deleted rab virus...
rabies is an rna virus... replicates only in cytoplasm...  glycoprotein produces  little red spikes...
 interacts with axon terminal 5 genes in rabies  NPMG and L...replace  G  with   EGFP...  will glow green
... stuck inside cell and cannot get out...
pseudotype:  express EnvA  difft enveope protein... will infect nothing...
express TVA  only in cells that we WANT to be  infected.
so use  deleted  G rabies as a retrograde tracer
infected cells  glow green.
label inputs to a single neuron in vivo
Jim Marshel did this work...
electroporate into a single cell in vivo
put in plasmids with TVA, YFP and rabies glycoprotein...
wait a few days...  expresses  3 genes...
spreads to presynap cells and turn  red  (with mCherry)  but they do not have glycoprotein so
that stops  the spreading of rabies...
also see distal inputs in vivo...
number of neurons... should get 500 to 1000 inputs  but only get max of 100
(ONLY synaptic inputs but not ALL the synap inputs)  see cells v far away that are infected
we label up to a 100... but we should be seeing 500 to 1000;
elucidating brain circuits in vivo
start with mouse line that express Cre recombinase
apply helper virus to express TVA
use AAV   flox construct...
protocol
AAV takes a few weeks to get AAV infection then
3 weeks later put in rabies...
cerebellar Purkinjes...
rabies virus can spread across parallel and climbing fibers
 in presence of Cre-dep glycoprtoein...
all work in mice...
PV= parvalbumin cells...  monosynap retrograde label...
thalamic cells in Postmed nucleus ... go to PV cells...
PV cells may be exclusive recipients of thalamic input
want to link struc to fxn...
rabies will kill cells after abt 2 weeks... meanwhile can do functional studies
put in GCamp3
put in ChR2
inject into  S1  barrel cortex...
  vision
can image down more than a mm (below pia) into extrastriate visual cortex...
while showing mouse visual gratings...
link connectivity to func properties...
 
 </p>