Escrever um comentário

Summary:

This important study by Possa-Paranhos and colleagues identifies a circuit-level mechanism underlying the phenomenon of hyperphagia (increased feeding behavior) in pregnant mice. First, using devices to continuously and precisely monitor feeding behaviors, the authors demonstrate that hyperphagia during pregnancy is due to an increase in meal size rather than meal frequency. Next, investigating neuronal populations in the arcuate nucleus of the hypothalamus, a critical hub controlling energy homeostasis, the authors find that pregnancy increases AgRP mRNA expression and the activity of AgRP neurons, while decreasing POMC mRNA expression and the activity of POMC neurons. Loss-of-function experiments using chemogenetics demonstrate that the inhibition of AgRP neurons and the excitation of POMC neurons both reverse hyperphagia in pregnant mice. Lastly, the authors utilize spatial transcriptomics to identify a striking number of pregnancy-induced transcriptomic changes in AgRP and POMC neurons that could underlie the changes in neuronal activity.

These results provide significant progress in the field’s understanding of how the brain adapts during the peripartum period to drive behaviors that support a healthy pregnancy and postpartum period. Given the inconsistency of prior studies and that the circuit-level mechanisms of this behavior had not been described, these findings fill a major gap in our understanding of the brain of pregnant mice. While some of the histology results would benefit from additional image analyses and the incorporation of behavioral data, the results overall support the authors’ conclusions that pregnancy induces an increase in AgRP neuronal activity and a decrease in POMC activity, each of which is necessary for hyperphagia. The consistency of the results across multiple approaches (mRNA quantification, fiber photometry, chemogenetics, pharmacology, and spatial transcriptomics) provides convincing evidence that the AgRP/POMC system adapts during pregnancy and drives hyperphagia.

Major points:

-       Throughout Figure 2, the representative images appear to show not only differences in intensity signal, but also significant differences in the number of AgRP and POMC neurons. With this in consideration, the RNAscope images and quantification require more analyses or more explanations to fully support the authors’ claim that pregnancy induces an increase in AgRP mRNA and neuronal activity and a reduction in POMC mRNA and neuronal activity. While I understand that it may not be the case that the number of AgRP and POMC neurons are changing during pregnancy (and perhaps it only appears this way because of a wide range of signal intensity between images), it would benefit the manuscript to explain this or to include, for example: number of AgRP/POMC neurons counted (area-normalized); mean signal or number of puncta per positive cell. Whether fasting and/or pregnancy are changing the number of neurons that express AgRP and POMC, or merely changing the expression within cells, is important for interpreting the rest of the manuscript. The number of AgRP and POMC cells counted also influences the interpretation of the cFos data – if the denominators (total AgRP/POMC) of the cFos fractions are significantly different between groups, this could be confounding the cFos percentages. To address this, the authors could (similarly to above) report some of the raw data or state in the text that the number of AgRP/POMC cells was not statistically different between groups.

-       To fully understand and interpret the cFos results, it would be very helpful for the authors to provide more details on the timeline of this experiment. Specifically, due to the nature of cFos dynamics, I am wondering if the mice were sacrificed relative to a feeding event or agnostic to the last feeding, since I imagine that feeding (and therefore hunger/satiety state) would alter the ratio of active AgRP and POMC neurons. Since the figure timeline indicates that feeding data was collected for these mice, the incorporation of this data into the analyses could be very informative and beneficial to the manuscript. For example, the authors could either normalize the cFos data to, or correlate the cFos data with, behavioral metrics such as: time elapsed between last feeding event and sacrifice; total amount of food eaten in the 30/60/90 minutes prior to sacrifice.

Minor points:

-       Figure 2:

o   The rationale for euthanizing the pups of the postpartum group is not clear and would benefit from more explanation. Since AgRP and POMC neurons are tied to the HPA axis, and the removal of pups is a stressor to which this group is uniquely exposed, I am unsure what conclusions can be drawn about the postpartum group. Moreover, since there is no postpartum behavioral data and this group only appears in Figure 2, the removal of the postpartum group data altogether could improve the manuscript by keeping the focus on pregnancy.

-       Figure 3:

o   To improve the interpretation of the data, it would benefit the reader to have more visual cues throughout this figure differentiating the longitudinal within-mouse comparisons, and the group comparisons. The longitudinal data is particularly compelling and could be highlighted more! One way to address this would be to add lines connecting datapoints from the same mouse across time.

-              Figure 4:

o   It appears that the mCherry control data for the AgRP-cre experiment is missing (after subfigure F; assuming that subfigures M-O are the POMC-cre mice injected with the mCherry control virus).

o   For consistency, since you show cFos/POMC colabeling increasing in hM3Dq mice after CNO, do you also have this data for the hM4Di experiment showing that CNO leads to a decrease in cFos/AgRP colabeling?

-              Figure 5:

o   Is there a difference in MC4R mRNA expression across groups that could explain or be contributing to the differential responses to the MC4R agonist drug (Figure 5)? Similarly to the above comments on Figure 2, it would be helpful to see quantifications of positive cells and expression per positive cell, since pregnancy-induced differences in MC4 receptors could also contribute to differences in MC4R agonist sensitivity (in addition to differences in melanocortin tone stemming from the AgRP and POMC neurons).

The reviewer’s expertise lies primarily in rodent behavioral and systems neuroscience. Therefore, the reviewer is unable to provide as much feedback regarding the transcriptomic experiments in this manuscript.

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.