Critical Commentary on "Gut-to-brain signaling restricts dietary protein intake during recovery from catabolic states" in Cell 2025

1 Critical Commentary on “Gut-to-brain signaling restricts dietary protein intake during recovery from catabolic states”in Cell 2025 Authors: Yiheng Wang and Shu-Feng Zhou* College of Chemical Engineering, Huaqiao University, Xiamen, China *Correspondence: [email protected] INTRODUCTION The study by Jaschke NP, et al.1 published in Cell (Nov 4, 2025) claims to uncover a gutto-brain signaling pathway that selectively suppresses dietary protein intake during recovery from catabolic states. According to the authors, animals recovering from fasting, inflammation, or glucocorticoid-induced catabolism exhibit a transient protein avoidance governed by a specialized enteroendocrine-vagal-NTS-PVN pathway. The authors propose that this mechanism protects the organism from “metabolic stress” imposed by early protein refeeding. The concept is striking because it contradicts long-standing physiological knowledge. Classical nutrition studies show that post-catabolic organisms typically demonstrate a protein appetite, seeking nitrogen replenishment and amino acids required for tissue re-synthesis. In contrast, this paper claims that protein intake is restricted during recovery, and that this restriction is actively imposed by a specific gut–brain circuit. Given its conceptual boldness, the claims require exceptionally rigorous evidence. Yet, upon detailed evaluation of the data across main figures, extended data, and supplementary figures, we find substantial methodological weaknesses, interpretive leaps, and missing controls. Several datasets support a general suppression of feeding rather than protein-specific avoidance. Circuit specificity is inadequately demonstrated. The purported gut-derived peptide remains unidentified and uncharacterized. The physiological rationale is speculative and inconsistent with established catabolic recovery literature. This commentary provides a comprehensive, figure-by-figure critique, evaluating technical and conceptual shortcomings that collectively undermine the strength of the authors’ conclusions. CONCEPTUAL CRITIQUE 2 Before examining individual figures, several overarching issues limit the interpretability of the entire study: 1. Protein avoidance vs. general hypophagia Many of the authors’ behavioral results show reduced total caloric intake rather than a selective avoidance of protein. Without detailed palatability, texture, caloric density, and sickness-behavior controls, differences in protein vs. carbohydrate intake may be driven by general malaise rather than nutrient-specific signaling. 2. “Catabolic states” are treated as interchangeable The authors employ several unrelated models: • 24–48 hr fasting • Glucocorticoid treatment • Hindlimb unloading • LPS-induced systemic inflammation These models share little mechanistic overlap. Grouping them under a single category (“catabolic states”) oversimplifies complex metabolic and immunological changes. 3. Gut-derived signal is not identified Despite claims of a gut peptide mediating the effect: • No peptide is purified • No receptor is identified • No cell-type–specific sensory mechanism is demonstrated Without molecular identity, the central mechanistic claim remains speculative. 4. Vagal and central circuits are nonspecifically labeled 3 The authors rely heavily on Phox2b-Cre, AAV, c-Fos, and PRV tracing—all of which suffer from: • broad tropism • multisynaptic contamination • lack of nutrient specificity • absence of electrophysiological validation Circuit specificity is therefore unsubstantiated. 5. Physiological rationale is inconsistent with known biology Protein restriction during recovery from catabolism contradicts: • nitrogen-balance literature • FGF21-driven protein appetite • classical starvation-refeeding physiology No biochemical or metabolic justification is provided. FIGURE-BY-FIGURE CRITIQUE (MAIN FIGURES 1–7) Below we critique each main figure individually. Figure 1 – Behavioral Evidence for Protein Avoidance Authors’claim: During recovery from catabolic states, animals avoid protein but not carbohydrates. Critique: 1. Total caloric intake drops, suggesting generalized hypophagia. 2. Palatability differences are not controlled; protein pellets usually differ in texture, hardness, and odor. 3. Circadian feeding patterns are not matched. 4 4. Lack of baseline preference comparisons before catabolic induction. 5. No microstructure licking analysis to distinguish taste vs. post-ingestive effects. 6. Some catabolic models (e.g., LPS) induce sickness behavior, confounding preference results. Conclusion: Evidence for selective protein avoidance is weak and confounded. Figure 2 – Identification of a Putative Gut Sensor Population Authors’claim: A subset of duodenal enteroendocrine cells (EECs) is activated by protein intake during catabolic recovery. Critique: 1. c-Fos staining in gut epithelium is unreliable, confounded by mechanical stimulation of feeding. 2. scRNA-seq is underpowered (fewer than ~400 EECs analyzed), resulting in unstable clustering. 3. No single-amino-acid sensing data—no leucine, arginine, alanine specificity assays. 4. No in vitro demonstration that these cells directly respond to proteins or amino acids. 5. No genetic markers distinguish this “protein-responsive subtype” from known EEC classes. Conclusion: Data do not convincingly identify a nutritionally specialized EEC population. Figure 3 – Vagal Mediation of Protein Avoidance Authors’ claim: Vagal afferents are required to transmit the gut-derived protein-avoidance signal. Critique: 1. Vagotomy causes major physiological disturbances (gastric emptying, motility changes) that confound nutrient choice assays. 5 2. Phox2b-Cre expression is not selective to vagal afferents. 3. Chemogenetic manipulation lacks validation (no immunostaining, no electrophysiology). 4. Behavioral rescue experiments confounded: CNO doses inconsistent, order of preference testing unclear. 5. No direct demonstration that vagal fibers themselves respond to protein stimuli. Conclusion: Vagal involvement is plausible but not demonstrated with specificity. Figure 4 – Hypothalamic Integration Node (PVN) Authors’ claim: PVN neurons integrate vagal information and suppress protein intake. Critique: 1. PVN c-Fos activation is non-specific; PVN responds to stress, novelty, and osmolarity changes. 2. PRV tracing is multisynaptic and prone to non-specific infection. 3. Optogenetic activation suppresses all feeding, not just protein intake. 4. No electrophysiological evidence for vagal-to-PVN monosynaptic connectivity. 5. No cell-type identification (e.g., CRH+, TRH+, oxytocin neurons). Conclusion: PVN recruitment is not nutrient-specific and is overinterpreted. Figure 5 – Secreted Gut-Derived Signal Authors’ claim: A secreted peptide from catabolic EECs suppresses protein intake. Critique: 1. No peptide is identified—LC–MS/MS data not adequate; no purification or receptor assays performed. 2. Conditioned medium contains many signaling molecules; the effect may be nonspecific. 3. AAV overexpression of the candidate factor lacks quantification and localization. 6 4. No demonstration that the conditioned medium activates vagal neurons. 5. No dose-response curves, time courses, or in vivo peptide injections. Conclusion: Major mechanistic gap; the gut-derived signal remains unknown. Figure 6 – Neurophysiology of the Vagal–NTS Pathway Authors’ claim: Protein ingestion activates a specific NTS population via vagal input. Critique: 1. Vagal fiber recordings lack amino acid specificity; no single-AA tuning curves. 2. Mechanical distension effects not controlled. 3. Calcium imaging in NTS prone to motion artifacts; insufficient controls. 4. NTS activation overlaps with GLP-1R+ cells—nonspecific satiety circuitry. 5. No causal link to protein suppression demonstrated. Conclusion: NTS activation is not protein-specific. Figure 7 – Circuit Manipulation (Opto/Chemo) Alters Feeding Authors’ claim: Activating or inhibiting the circuit modulates protein intake. Critique: 1. Optogenetic activation of PVN reduces all feeding, not selectively protein feeding. 2. No carbohydrate/fat intake data under identical manipulations. 3. Locomotor, anxiety, and sickness-behavior controls insufficient. 4. Inhibitory manipulations show inconsistent behavioral effects. 5. Viral expression heterogeneity not quantified. Conclusion: Manipulation results consistent with global satiety, not protein-specific regulation. 7 EXTENDED DATA FIGURES ED1–ED15 Below we provide a concise critique of each Extended Data figure. ED1 – Validation of Catabolic Models • Models are physiologically distinct and cannot be grouped together. • No amino-acid profiling, nitrogen balance, or corticosterone levels measured. ED2 – Palatability Controls • Only minimal lickometer data; no texture, hardness, or energy density matching. • Protein pellets visibly different. ED3 – Organoid EEC Characterization • EEC differentiation markers incomplete; organoids lack physiological context. ED4 – scRNA-seq Clustering • High mitochondrial gene percentages; clusters unstable; batch effects unaddressed. ED5 – Vagal Fiber Labeling • Phox2b-Cre not selective; tracing overly broad. ED6 – Amino-Acid Stimulation Panel • Only AA mixtures tested; no individual amino acid tests. 8 ED7 – NTS Neuropeptide Expression • Bulk qPCR insufficient; no cell-type resolution. ED8 – Optogenetic Validation • No electrophysiological confirmation of opsin functionality. ED9 – Behavioral Control Assays • Activity scoring underpowered; sickness behavior not quantified. ED10 – Short-Term “Chronic” Manipulations • 48 hours too short to evaluate catabolic recovery. • No metabolic, endocrine, or nitrogen-balance correlates. ED11 – Hormonal Profiling • Missing key catabolic hormones (FGF21, IGF-1, GH, urea). • Inconsistent sampling times. ED12 – Gut Morphology • Descriptive only; no quantification of villi, crypts, or EEC density. ED13 – Vagal Electrophysiology Controls • No raw signal metrics; spike sorting unclear. 9 ED14 – Viral Tool Specificity • AAV tropism broad; Cre-lines leaky; no colocalization quantification. ED15 – Longitudinal Behavior • No nitrogen balance; high variability; behavioral entrainment confounded. SUPPLEMENTARY FIGURES S1–S12 S1 – Diet Composition Tables • Amino acid profiles missing; fiber and moisture differ between diets. S2 – Additional Lickometer Data • Burst structure, satiation curves missing. S3 – scRNA-seq QC Metrics • Poor library complexity; excessive mitochondrial gene expression. S4 – Viral Controls • No inflammation scoring; GFP toxicity unaddressed. S5 – Additional NTS Imaging