PREreview del Common and distinct roles of AMPKγ isoforms in small-molecule activator-stimulated glucose uptake in mouse skeletal muscle

The review is the result of a collaborative live review discussion held on 4 November 2025, organized and hosted by the PhD School in Biomedical Sciences of the University of Padova as part of the Journal Club activity. The discussion was followed by additional asynchronous work by students and the JC committee, who jointly contributed to the composition and refinement of this report. This work was posted on bioRxiv on 29 September 2025 and now is published in Molecular Metabolism (https://www.sciencedirect.com/science/article/pii/S2212877825002017). Although the paper has already undergone full peer-review, we believe that our comments may help the authors and other researchers to identify gaps that could be addressed in future studies in the field.

The work by Biswas and colleagues aims to determine the regulatory role of the AMPKγ subunit in response to ADaM site-binding activators, hypothesizing that the γ1 isoform is necessary for ADaM site activator-stimulated glucose uptake in skeletal muscle. To test this, the authors used a combination of single-nucleus RNA sequencing (snRNA-seq) on mouse and human skeletal muscle samples, AMPKγ1/γ3 single and double knockout mouse models, ex vivo glucose uptake assays, and in vivo glucose measurements .

The work presents a linear and consistent workflow, from molecular to functional assays. The introduction is clear and informative, and the relevant state of the art is clearly depicted and referenced. We particularly appreciated the clear cell type–specific characterization of AMPK γ-isoforms, both in mice and humans, as well as the use of mouse models with isoform-selective knockouts.

While the data are generally consistent, we have some suggestions that could help the authors strengthen their conclusions:

Major comments:

● In figures 4 and 5, pT172 AMPK is quantified relative to vinculin. However, it would be informative to assess the total level of AMPK proteins and eventually the level of phosphoAMPK/totalAMPK ratio in order to have the quantification of the fraction of phosphorylated protein. The authors should integrate the analysis with this piece of information and re-interpret the results accordingly.

● While the authors present a robust model with the knockout of both AMPK subunits, there is an incomplete characterization of the mouse models’ phenotype, which is not described in sufficient detail. The paper lacks information regarding whether the loss of these subunits affects skeletal muscle morphology, fiber composition, or overall muscle mass and force. Moreover, given AMPK’s role in glycogen metabolism, an assessment of intramuscular glycogen levels in knockout models would be highly informative.

● As a mechanistic follow-up, we suggest that a knock-in model selectively disrupting the ADaM binding pocket (e.g., in AMPK β-subunits) could provide a more specific test of ADaM activator dependence than γ-subunit knockouts, which alter overall AMPK complex abundance/stability. An ADaM-site KI would help distinguish on-target ADaM signaling from secondary effects of reduced AMPK heterotrimer levels.

Minor comments:

● The quality of the UMAP figures is low, please upload images with a better definition.

Methods section

● In Figure 2, the authors state that n = 6 human samples were tested for snRNA-seq. However, in the methods section it is written that 9 male and 6 female samples were collected. The authors should clarify which samples were used for the analysis and discuss potential sex/age differences.

● The snRNA-seq analysis is technically strong and appropriately applied to skeletal muscle. However, skeletal muscle is composed of multinucleated myofibers, such that a single nucleus does not correspond to a single cell. Consequently, the relative abundance of nuclei in the dataset may not accurately reflect true cell-type proportions and can result in over-representation of myofiber-derived nuclei, introducing sampling bias when interpreting cell-type gene expression. While the authors do not explicitly equate nuclear fractions with cellular proportions in the results, the manuscript would be strengthened by an explicit statement acknowledging that nuclear counts do not directly translate to cell counts. In addition, the analysis would benefit from benchmarking cluster annotations and marker definitions against established skeletal muscle nuclear atlases, such as Petrany et al. (2020) and Dos Santos, et al. (2020).

Results section

Section 3.1

● The differences regarding the distribution of AMPKγ isoforms between mice and humans are only briefly described. The γ2 isoform is absent in mice but is clearly visible in humans. Could this be due to differences in the muscle types analyzed (gastrocnemius/vastus lateralis)? This discrepancy should be explained in more detail.

● As for the transcript levels analysis, we believe it is also important to check the levels of the other subunits: α and β.

Section 3.2

● The mRNA levels of AMPKγ2, are referenced to figures 1N and O, but the figures are not present in the present article or in the referenced article [11].

● We also suggest analyzing γ2 protein levels by Western blot.

● The destabilization of the α and β subunits seen in Western blot analysis of knockout sample would have experiment specific consequences, because many of the assays assume and intact AMPK heterotrimer. Could this issue influence the interpretation of some of the experiments performed?

For instance, the AMPK activity and phosphorylation assay, the reduction of p-AMPK (Thr172) and p-ACC could be structural, not regulatory. Within the small-molecule AMPK activation experiments where the apparent isoform selectivity could be confounded by the reduced α/β stability, where the reduced drug response might reflect loss of target, not resistance and this is relevant given the focus on γ-isoform specificity. In summary, the destabilization of the AMPK α and β subunits would most strongly affect kinase activity measurement, pharmacological activation experiments and functional metabolic assays, potentially confounding interpretation of γ-isoform-specific effects. We suggest that the authors mentioned that some phenotypes may arise from altered holoenzyme (complete and catalytically active form) stability rather than isoform-specific signaling alone.

● The statistical significance of the results is unclear in certain figures (Figures 4, 5 and 7) where the “#” symbol is often used. We suggest using the common notation with asterisks based on p-value significance, and clearly defining all symbols in the figure legends.

● Impairment of AICAR-stimulated glucose uptake mediated by the lack of the γ subunit was already observed in a previous study (Rhein et al., 2021). The authors should explain the novelty of their results regarding this aspect, particularly in the context of ADaM-site activators and isoform-specific requirements.

Discussion session

● The authors should further discuss the structural and mechanistic importance of γ1 and γ3, considering the data presented.

Conclusions section

● Conclusions are extremely brief, somewhat superficial, lacking mechanistic depth and thorough discussion of the results obtained. We suggest the authors expand this section.

Competing interests

The reviewers declare that they have no competing interests.

Reference:

1) Petrany, M.J., Swoboda, C.O., Sun, C. et al. Single-nucleus RNA-seq identifies transcriptional heterogeneity in multinucleated skeletal myofibers. Nat Commun 11, 6374 (2020). https://doi.org/10.1038/s41467-020-20063-w

2)  Dos Santos, M., Backer, S., Saintpierre, B. et al. Single-nucleus RNA-seq and FISH identify coordinated transcriptional activity in mammalian myofibers. Nat Commun 11, 5102 (2020). https://doi.org/10.1038/s41467-020-18789-8

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.