,
13 tweets,
7 min read
Read on Twitter
The GPe is important for motor control but its precise involvement is unclear. A major challenge has been the identification of neuron subtypes. Our lab along with others such as @ArynGittis and @kaabrahao @_pauldodson @blsabatini have been working hard on the topic. 1/n
Based on the earlier work from @MarinLab and others, we’ve made a Npas1 BAC mouse and two novel abs for Npas1. Importantly, we found that PV+ neurons and Npas1+ neurons are two near-exclusive (~2% overlap) neuron classes. 2/n
However, there are up to 20% of the GPe neurons that we have not been molecularly defined. Others labs arrived at different classification schemes. That itself is a whole other thread, for a different day. OTOH there are a few outstanding questions. 3/n
1) One major issue has been the discrepancies in the data across labs, e.g. PV+ neurons are estimated to be 30–60% of the GPe. That’s a big range! 2) Another debate has been the abundance of Lhx6+ neurons. But can they account for the double negative PV– and Npas1– neurons? 4/n
3) @blsabatini and @lindsayschwarz published on Ctx-projecting GPe neurons but they are a mixture of cholinergic and GABAergic neurons. While the former is convincingly demonstrated by Bernardo’s group, the identity of the latter was not as clear. 5/n
Are these questions all tied to ea other? We determined to reconcile these w/ the advent of new transgenic lines. Based on the work from @GordFishell and others, we decided to examine molecules that are upstream (i.e. Nkx2.1) and downstream (i.e. Sox6) of Lhx6. 6/n
We found Sox6 to be a new defining marker for GPe neurons. Lhx6 neurons that co-express PV or Npas1 also co-express Sox6. This provides a new, independent metric to demonstrate that there is indeed a unique Lhx6 neuron population (i.e. Lhx6+Sox6–). 7/n 
That’s the 1st key finding of this paper. As @MarinLab previously found neurons from the Dbx1+ lineage populates the GPe, we wanted to ask if they correspond to this Lhx6+PV–Npas1– population. The quick answer is no, instead, they give rise to mostly Sox6+ and PV+ neurons. 8/n
As the topography of the Ctx-GPe projection hasn’t been precisely mapped, we set out to examine that more closely. It turns out M1/S1 form a tight reciprocal feedback loop with the GPe. This finding has far-reaching implications on motor control and decision making! 9/n
To study the microarchitecture of the Ctx-GPe loop, we used immuno, transgenic, and patch-clamping approaches. We determined that Npas1+Nkx2.1+ neurons give rise to the GABAergic projections to Ctx (and reticular thalamus). 10/n
It is clear that the connection is sparse (<10%) but the connection strength can be quite large. We are still trying to make sense of our data but the connection probability can be up to 70% (or higher?). Importantly, they produced the biggest thing we have ever seen—3 nA! 11/n
We have developed a new method to visualize the abundance and distribution of GPe neuron subtypes. It’s a ton of work but it’s a fun and objective way to see how neuron subtypes are spatially distributed within the GPe. 12/n
By examining the molecular profile, axonal projections, and electrophysiological properties, we concluded that Npas1+Foxp2+ & Npas1+Nkx2.1+ neurons should be considered as two unique GPe neuron subclasses. 13/end
