Escribir un comentario

The authors express mWmC in a genetically defined manner to trace synaptic connectivity, starting with SST+ or NOS1+ neurons, and separate input neurons from target neurons based on reporter expression. In the first part of Figure 1 they use Baf53b-Cre to show that the tracer faithfully labels neurons and not other cell types, with labelling in both input and target carrying the same marker (mCherry). In the second part of Figure 1 they use Cck-Cre to show labelling in neurons in the gut traced to the celiac–superior mesenteric ganglia. In Figure 2 they use a lox-stop-lox nuclear eGFP reporter to label input cells and distinguish targets within the same tissue.

This preprint fills a genuine gap in the field by describing a tool to trace ENS connectivity. ENS connectivity work thus far has relied on electrophysiology, dye fills, or post hoc staining. This provides a change in throughput and more rigorously illustrates connectivity, as it shows connectivity via label transfer rather than inferring it from location or static markers.

The group found the study straightforward and the data internally consistent, with tight data points and impressive results. Notably, 100% of labelled cells being neurons and 0% glia. SOX10+ glia are not labelled despite glial contact with neurons, and c-KIT+ ICCs are likewise spared, so the tracer appears genuinely restricted to neurons. The inclusion of a time-resolution experiment and the introduction of the dual reporter system were both regarded as strengths. The plateau experiment is the right way to test monosynaptic restriction, as the tracer does not recruit more cells at successive timepoints; comparing day 10 and day 28 post injection shows no further spreading, which gives confidence that the label does not continue to travel and that true input/target relationships are being mapped.

The comments below are offered constructively, in the hope they are useful as the manuscript moves toward publication. Our major points concern the interpretation of preferential connectivity, demonstration of directionality in the ENS specifically, and the reporting of statistics; the minor points are largely presentational.

Major comments

1. Preferential connectivity is inferred without accounting for the relative abundance of each marker.

The authors state that preferential connectivity is mapped, but they measure this by comparing between markers in the target pool without considering whether the neuron numbers for those markers are equally abundant. For example, for somatostatin they state that it preferentially connects with CALB2 neurons because of 47% labelling, compared with NOS1 at 4% labelling. It is more informative to consider the total abundances of NOS1 and CALB2 neurons. The claims on SST to CALB2 are slightly overstated at this sample size (3–4 mice), since the CALB2 baseline is around 35%, so finding 47% is not far enough outside that to establish preference. Resolving SST to CALB2 at the paper's variability would require approximately 15 mice, and NOS1 to CALB2 approximately 6.

2. The NOS1 to CALB2 result may point in the opposite direction to the claim.

For NOS1 to CALB2, approximately 20% connectivity labelling is reported against a published baseline of about 35%. The authors describe a robust synaptic connection, but at this level of labelling the value sits below the published abundance and so trends more towards avoidance (although at this sample size that claim cannot be made either). We would suggest softening the current wording accordingly.

3. A within-animal enrichment analysis would address both points without additional mice.

The authors could score the same markers across all HuC/D+ neurons in the same ganglia, and compute the percentage of target cells positive for a marker divided by the percentage of all local neurons positive for that marker, testing the resulting ratio against 1. Performing this within animal would not require the higher sample sizes noted above.

4. The strongest result is currently underemphasised.

The preprint shows convincingly that somatostatin neurons essentially do not contact nitrergic neurons. This deserves greater emphasis, because it is the claim best supported relative to published baselines: 4% labelling is reported, against roughly 33–39% NOS1 neurons in mouse colon or 26% in rat ileum, and so falls far outside the expected range.

5. Directionality should be demonstrated in the ENS rather than inherited.

The parent Trans-Seq paper notes that unfused WGA protein is bidirectional, and the engineering of mWmC with anterograde versus retrograde quantification shows limited retrograde spread, but this was established in retinal and cortico-striatal circuits. The ENS has dense connections, so we would encourage the authors to demonstrate this in the present context rather than assume it carries over.

6. No statistics are reported.

There are no reported statistics throughout, and including them would strengthen the preprint considerably.

7. Within-class connectivity cannot be assessed, and this belongs in the main text.

The construct is very useful for between-class connectivity, but it is not possible to examine within-class connectivity because the tool is genetically based; SST to SST connectivity, for example, cannot be detected. The authors make this point in the legend to Figure 2G (“this neuron class could not be evaluated as a potential target neuron as it cannot be distinguished from input neurons”), but it is worth stating in the main text as a limitation of the approach.

8. Non-toxicity is asserted but not measured.

The authors state that there is no toxicity, but no measurements are shown to support this. Caspase-3 staining or neuron counts between mice would help substantiate the claim. More broadly, the group wondered whether introducing this construct influences cells in other ways (i.e. cell health or function) and some evidence on this point would be reassuring.

9. Language describing the order of transfer could be made consistent.

The claim is variously framed as monosynaptic and as first-order transfer. We would suggest settling on one formulation and using it consistently throughout.

Minor comments

• Line 58: the definition of the ENS would be more accurate as interconnected neuronal subtypes and non-neuronal cells, so that glia are explicitly included.

• Summary: it would help to explain the relationship of the celiac–superior mesenteric ganglia to the gut, if room permits.

• Line 85 (Figure 1D): no overlap between c-KIT and mCherry is described, but this is hard to judge at the magnification shown. The broad overview is useful; a higher-magnification inset would make the absence of overlap convincing.

• Figure 1L: similarly, the absence of overlap would benefit from clearer presentation.

• Is there a particular reason for focusing on the distal ileum? A sentence of rationale would be helpful.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.