PREreview of The gut-brain vagal axis governs mesolimbic dopamine dynamics and reward events

Summary

Using a subdiaphragmatic vagotomy (SDV) model to perturb gut-brain communication, here Onimus et al. investigate the contribution of the gut-brain vagal axis to reward-mediated behaviors and their underlying dopaminergic mechanisms. They first establish the importance of an intact vagus nerve for both natural and drug-reward seeking behaviors across a suite of behavioral tasks, the range of which allows the authors to be impressively specific about the processes affected by this manipulation. They then reveal for the first time that the SDV manipulation alters the dynamics of dopamine release and molecular activity of neurons in the nucleus accumbens (NAc) to both food and drug reward, directly implicating the mesolimbic dopamine system. Finally, the authors probe the upstream drivers of these changes, finding that the spontaneous activity of ventral tegmental area (VTA) dopamine neurons is altered via reduced glutamatergic input. This results in changes in receptor dynamics and dendritic spine density downstream in the NAc but not the dorsal striatum. Their approach of systematically probing multiple nodes in the vagus-mesolimbic dopamine network allows the authors to build a preliminary mechanistic model of this circuit and its involvement in behavior.

Relative to the amount of work done on external reward processing, the field’s understanding of how interoceptive signals shape mesolimbic dopamine signalling is limited. The findings of this study are therefore important to furthering an integrative understanding of a crucial brain-body pathway and its role in behavior. However, although the authors deeply interrogate multiple nodes in the system through use of a broad range of behavioral, physiological, and molecular techniques, this convincing dataset could be made compelling by further assessing the specificity and endogenous control of vagal afferents onto VTA dopamine neurons. 

Major comments 

1.        The authors’ working model appears to be that of a vagus - NTS - PBN - VTA circuit. But there are two unresolved issues that have to be addressed to verify this as a full model of vagal gating of mesolimbic dopamine reward processing:

a.        As the authors point out in the discussion, whether the altered excitability of VTA dopamine neurons is through glutamatergic input within the VTA microcircuit or from the PBN is unanswered here. An alternative is that the glutamatergic inputs actually originate from another region also innervated by the NTS and that are known to synapse onto neurons in the VTA, such as the locus coeruleus, dorsal raphe, or hypothalamus. 

b.        The current study focuses on the effect of constitutive activity of the gut-brain vagal axis. However, endogenously this input can act on a range of timescales, from seconds to hours. The SDV manipulation is chronic and thus the specific effects seen on D2 receptors in the NAc (namely, changes in excitability) are presumably a consequence of this longitudinal timescale of the manipulation, but might not be reflective of the endogenous function of vagal ascending input into this system.

To resolve this, the authors should image PBN axons in the VTA in order to compare endogenous responses to their SDV model. If this is beyond the scope of the current work, it would be helpful for the authors to outline these other possible circuit models and this limitation in greater detail in the discussion.

2.        The authors begin by characterizing differences in mouse behavior in the SDV model relative to sham controls in response to natural rewards. They do so using a laudably wide range of behavioral tasks as well as parallel controls for e.g. metabolic states in order to probe different aspects of behavior but also of cognition - including motivation, learning, ‘liking’, incentive salience, reinforcement, and conditioning. But the authors’ final interpretation is confusing; for example, they conclude that this gut-brain interaction is necessary for ‘positive reinforcement and conditioning’ but earlier say that ‘learning is intact’. In reality, these tasks each tap into multiple cognitive processes (e.g. the t-maze involves learning, motivation, reinforcement). A clearer integration of the results across these tasks would be beneficial for understanding the role of this pathway. To be more specific, the effects seem to be predominantly on the motivational axis - which would also be consistent with the changes in excitability/temporal scale of the SDV manipulation - so making this explicit (or discussing otherwise) would be helpful.

3.        As the authors mention in the introduction, natural and drug rewards share some mechanistic properties, but not all. Here the authors’ conclusion seems to be that the vagus nerve to the mesolimbic dopamine system is a common pathway for the behavioral effects of both reward types. However, there are a number of differences in the response of SDV animals to natural and drug rewards. For example, only anticipatory locomotion is reduced in response to the palatable solution, while all locomotor activity is affected by dopamine-mediating drugs. Similarly, dynamics in the NAc in response to natural reward in SDV animals appears to be almost entirely ablated, while with drug reward it is more of a delay. I therefore suggest including something about possible differences in the vagus-axis mesolimbic dopamine encoding of natural and drug rewards in the discussion. 

Minor comments

1. As a synthesis of the authors’ results, I suggest including a schematic as a final figure to be referenced within the discussion. This will help clarify the authors’ working model of the system and will be a useful reference point for readers. 

2. For visualization of traces please make it so that one trace doesn't preclude being able to see the other - this is especially an issue in figures 3B and 3C.

3. The use of acronyms is inconsistent throughout the paper; the PS acronym is given but then not used, so can probably be removed entirely. Please spell out NTS and PBN in the results section, whenever it is first used, in order to improve reader understanding. And please use NAc, not Nac, consistently. 

4. By the dosages given it is implicit that all drug applications were intraperitoneal, but please make this explicit in the methods for clarity.

5. This is entirely the authors’ preference, but I suggest a slight rewording of the title - it is unclear what the gut-brain vagal axis governing ‘reward events’ means. Maybe something more like ‘processing of reward events through the mesolimbic dopamine system’ as an alternative (but again, just a suggestion). 

This reviewer states no competing interests. The reviewer wishes to remain anonymous, but their expertise primarily lies in the fields of rodent behavioural and systems neuroscience.  Thus, they are less able to evaluate some of the molecular level work contained in this manuscript.

AI was not used in the generation or editing of this review.

Competing interests

The authors declare that they have no competing interests.