The unique synaptic circuitry of specialized olfactory glomeruli in Drosophila melanogaster

Introduction to untangling pasta (neural connectomics)

Neuroscience borrows a lot of terminology from electrical engineering, like “circuit motifs” and “wiring diagrams.” Words like these suggest straight lines—with inputs connecting to outputs in an orderly manner. But in reality, terms like “pasta plate” or “spaghetti diagram” would be more apt for describing how neurons are connected in the brain of living organisms. The brain is wetware, not hardware. A recent study by Gruber and colleagues uses laser branding and electron microscopy to try to make sense of this welter of living wires. The authors investigated the structure of the olfactory regions in the Drosophila melanogaster brain that respond to specific odors, as compared with those that respond to many odors.

In insects (and similarly in vertebrates), olfactory sensory neurons on the antennae converge in a region called the antennal lobe. Within this region, neurons detecting a particular odorant or group of odorants form spherical clusters known as glomeruli, which literally means “little balls of thread.” Projection neurons transmit information from glomeruli to higher brain centers. Local interneurons connect glomeruli together, allowing them to modulate each other’s activity.

That was the simple wiring diagram description. Now for the details that make this more like a spaghetti plate: neurons can synapse onto themselves; most neuronal outputs connect to multiple inputs (termed polyadic synapses, see Fig. 1 of the preprint); dendrites responsible for input processing also have outputs (dendro-dendritic synapses); and, finally, a whole smorgasbord of neurotransmitters and neuromodulators add additional subtlety and complexity—akin to the seasoning on pasta.

Gruber and colleagues wondered if glomeruli that respond to only a few odors have common features that differ from glomeruli that respond to many odors. They used focused ion beam-scanning electron microscopy to investigate the DA2 glomerulus, which is narrowly tuned and responds to the highly aversive odorant geosmin, and the DL5 glomerulus, which is broadly tuned to a number of aversive odorants such as benzaldehyde (I keep them straight by remembering that the one with the bigger number responds to more odorants). The researchers added VA1v to their comparison, a narrowly tuned glomerulus that responds to the aggregation signal methyl laurate. The scientists took images at 20 nanometer distance intervals (unimaginably small steps) via focused ion beam-scanning electron microscopy, and they then reconstructed 3D volumes via computer software. To keep track of the location of their glomeruli of interest, the authors used a two-photon laser to brand lines in the brain around the glomeruli—thereby making it easier to find the edges of the glomeruli in all those virtual slices.

 

Fig. 1. Overview of the glomerular structure. A) and B) both show the two main glomeruli investigated in this study via in vivo functional imaging. The thin green line above and below both glomeruli is the laser branding mark. C) and D) show these glomeruli as scanning electron microscope images. In D), the white triangle in the bottom right points to a thin black line that corresponds to the laser branding mark. E) shows an example of a polyadic synapse (a tetrad in this case), both as a scanning electron microscope image and as a diagram. F) shows a digital reconstruction of an olfactory sensory neuron based on the electron microscope images.

Highlighted results of the study

This study finds a number of interesting patterns that could lay the groundwork for future hypotheses. The authors distill their comparison of broadly tuned versus narrowly tuned glomeruli into five main findings:

1. Narrowly tuned glomeruli have stronger feedforward olfactory sensory neuron connections to both uniglomerular projection neurons and multiglomerular neurons. That is to say, the input neurons have more connections to output neurons and interneurons.
2. Narrowly tuned glomeruli have much stronger axo-axonic communication between sister olfactory sensory neurons. In other words, the input neurons have more possibilities to modulate each other’s input.
3. Narrowly tuned glomeruli have stronger dendro-dendritic connections between uniglomerular projection neuron dendrites. The output neurons are better able to synchronize with each other and modulate each other.
4. Narrowly tuned glomeruli have less feedback from uniglomerular projection neurons to multiglomerular neurons, a group that includes both multiglomerular projection neurons and local interneurons. The output neurons do not have as many opportunities to modulate the neurons that communicate between glomeruli.
5. Narrowly tuned glomeruli have less feedback from multiglomerular neurons to olfactory sensory neurons. There are fewer opportunities for interneurons and multiglomerular projection neurons to modulate the input neurons, i.e., the inputs are subject to less “cross-talk” from other glomeruli.

These findings support the notion that narrowly tuned glomeruli establish more direct links with higher brain centers, with inputs being less influenced by other glomeruli. Conversely, broadly tuned glomeruli seem more receptive to signals from neighboring glomeruli. Narrowly tuned glomeruli exhibit greater intraglomerular modulation of inputs and outputs, potentially amplifying and synchronizing signal transmission.

One final finding that stands out: autapses, or synapses of a neuron onto itself, occurred at a greater frequency than the authors expected. This finding was especially true for the one uniglomerular projection neuron that leaves the broadly tuned DL5 glomerulus. It appears that the large, sprawling dendrite of this neuron synapses onto itself such that distant regions are more interconnected—perhaps providing the possibility for synchronizing the activity of the neuron.

Why I liked this preprint

As big, whole-brain connectomes become increasingly feasible, the methods used in this study for quickly making connectomes of important regions become especially interesting—they could make it possible to compare select regions across multiple individuals, to see just how representative one connectome can be. Moreover, because these methods link connectomics to functional work, it could be possible to compare between individuals that respond differently to a given stimulus and see how the physical connectivity of the brain influences different responses. Finally, this study provides a lot of interesting basic information about how important olfactory pathways are set up.

Questions for the authors

1. In the study, you use the category “multiglomerular neurons” for both local interneurons and multiglomerular projection neurons. Could you elaborate on why you had to put these two disparate types of neurons in the same category, while you were able to differentiate olfactory sensory neurons and uniglomerular projection neurons? What additional steps would it take differentiate these?
2. Do you have any reason to think that there might be similar patterns of differences in the multiglomerular projection neurons as with the uniglomerular projection neurons between broadly and narrowly tuned glomeruli?
3. Do you think your method could complement whole connectomes of a single brain by comparing a region of interest in several individuals, or do you see this method as opening up other avenues of research?
4. In your workflow, which parts of the process actually take the most time and are the main bottleneck? The tracing perhaps?
5. In Figure 2 – figure supplement 1, it looks like there are some synapses without connections (no postsynaptic contacts per T-bar). Could you elaborate on that finding? Are those synapses in the process of forming/being pruned, or could they be the product of damaged brain slices?

Tags: fly, fruit fly, insect, neural circuits, neuronal architecture, olfaction, smell

doi: https://doi.org/10.1242/prelights.35133

Read preprint (No Ratings Yet)