Lower motor neurons have long been understood to be particularly vulnerable to diseases like amyotrophic lateral sclerosis (ALS) and spinal muscular atrophy (SMA). These spinal cord neurons are cholinergic, meaning they use acetylcholine communicate with the muscles they control. We know that certain subtypes of these neurons die in these diseases while others are spared, but we don’t understand why. What is it about subtypes of these spinal cholinergic neurons that makes them more susceptible or resilient to disease? That’s the question that drove this project ¹. 

One of the biggest gaps in knowledge that makes it difficult for us to answer this question has been the field’s limited understanding of spinal cholinergic neuron subtypes. With the advent of single cell RNA sequencing technologies, it became possible to ask what subtypes actually exist as defined by their transcriptomes, and how distinct subtypes of these neurons differ from one another. Still, several studies that sequenced spinal cord cells had a striking underrepresentation of cholinergic neurons – possibly because of their relative rarity (< 1% of spinal cord cells) or their larger size (making them technically more difficult to isolate). We overcame this challenge by taking by selectively enriching for cholinergic neurons using a strategy in which GFP fused to the nuclear envelope protein Sun1 is Cre-dependently expressed. This allowed us to selectively tag nuclei of cells expressing the cholinergic neuron marker Chat, and isolate them with fluorescence-activated cell sorting for single nucleus RNA sequencing.

With the help of our collaborators at NIH, we successfully sequenced thousands of adult spinal cholinergic neurons. Excited by our success, my advisor, Dr. Claire Le Pichon, decided to include a teaser of this data at the end of her presentation at the 2020 Packard Meeting, an annual meeting of ALS researchers from around the globe. This is how we found out that Dr. Aaron Gitler’s lab at Stanford had been generating very similar data using an analogous method and was preparing to submit their manuscript. We decided to complete our analysis and write a paper as fast as we could – just in time for a global pandemic. 

As a bench scientist, being stuck at home was the perfect time to get better acquainted with R and dive into sequencing data analysis. I was fully immersed in the data and learned more about cholinergic neurons than I’d ever known, including interneurons and preganglionic neurons in addition to the skeletal MNs that had motivated the work. Although much of our goal was to identify more specific markers for cholinergic neuron subtypes involved in disease, we made many other unexpected discoveries along the way. Our dataset uncovered an unexpected spatially restricted diversity of preganglionic visceral motor neurons and a large population of these neurons in the cervical spinal cord. These autonomic neurons are canonically found in thoracic and sacral levels, not cervical. We don’t yet know what functions these cervical-level preganglionic neurons are responsible for, but we were able to confirm their presence. We also found spatially restricted alpha motor neurons, and validated combinatorial markers for digit-innervating motor neurons and the diaphragm-innervating phrenic motor neurons. Phrenic motor neurons are highly resistant to degeneration in ALS and, if preserved, could keep an ALS patient alive. The identification of markers for phrenic motor neurons is a significant advance for their study.

The data collected in this study provides an incredible amount of information, but we can’t investigate it all.  We also uncovered a third type of skeletal motor neuron that likely represents beta motor neurons. This dataset provides information that may finally enable a full characterization of this elusive cell type, for which an exact function is unknown. We found 16 types of visceral motor neurons that varied along the length of the spinal cord; now the question is which autonomic ganglion each of these subtypes innervates, and why preganglionic motor neurons have such different transcriptional identities. Beyond motor neurons, we found 8 types of cholinergic interneurons, of which only one type, partition cells, had previously been described; the next step is to understand what they are and their functions.

 We could not have anticipated that Claire’s last-minute addition of a couple slides to her presentation at a conference would lead to a race to publish, and eventually, the integration of our data with the Gitler group’s data, now publicly available on an interactive website (http://spinalcordatlas.org/). These datasets represent an incredible advancement in the study of adult spinal cholinergic neurons. I am exceedingly proud of this work and the effort that went into it, and I am eager to see future studies that result from it.

Our data, integrated with the Gitler group’s data, at spinalcordatlas.org (thank you, Jacob Blum). Our data by itself, as well as all future sequencing data from the Le Pichon lab, will be available at seqseek.ninds.nih.gov (thank you, Dr. Ariel Levine).

1. Alkaslasi, M. R. et al. Single nucleus RNA-sequencing defines unexpected diversity of cholinergic neuron types in the adult mouse spinal cord. Nat Commun 12, 2471 (2021).