Author response:
The following is the authors’ response to the original reviews.
eLife Assessment
This study presents a useful finding on the effects of arginine vasopressin (AVP) on islet cells in pancreatic tissue slices, using technically sophisticated spatio-temporal calcium recordings to confirm that AVP influences α and β cells differently depending on glucose concentrations. While the study's methods - particularly the calcium imaging techniques and peptide ligand design targeting V1b receptors - are strong, the reviewers were concerned about several aspects of the experimental design. However, the results on βcell responses are incomplete and insufficient to support the manuscript's claims, especially due to the high variability of islet responses and lack of mechanistic and functional (hormone release) data. There are also concerns about the possibility of off-target effects and incomplete receptor specificity, noting that the study would have been significantly strengthened by inclusion of signaling pathway interrogation, hormone output assays, genetic validation (e.g., β cell-specific deletion of V1br), and receptor localization, although the work will still be of interest to researchers studying islet physiology in the context of health and diabetes.
We sincerely thank the reviewers and editors for their thorough evaluation of our manuscript and their recognition of its technical strengths, including the advanced spatio-temporal calcium imaging and the rational design of selective V1b receptor ligands. We appreciate their acknowledgement of the study’s relevance for understanding AVP effects in a physiologically intact islet context and their positive assessment of our methodological rigour and innovation. The reviewers’ constructive feedback has helped us clarify the boundaries and intent of our study, which focuses on the glucose- and context-dependent modulation of α- and β-cell activity, rather than exhaustive molecular dissection.
While the reviewers rightly emphasize the importance of receptor specificity and downstream signaling validation, we respectfully suggest that some of their concerns may reflect a lingering bias toward reductionist frameworks. Our interpretation is rooted in the emerging understanding that β-cell behaviour is largely defined by dynamic intercellular interactions within the islet collective, rather than by static gene expression or receptor localization alone (Jin et al., 2025; Korošak et al., 2021; Rutter et al., 2024). Recent studies have demonstrated that roles such as “leader” or “hub” β cells are transient and emergent, governed more by timing, environment, and local network structure than by fixed molecular identity (Postic et al., 2023; Gosak et al., 2018).
This has profound implications for how we interpret cell responsiveness to agents like AVP: what appears as biological variability may in fact reflect context-sensitive transitions within a non-linear, self-organizing system (Stožer et al., 2021). Hence, we chose to focus on functional collective dynamics using intact pancreatic slices, rather than isolated cell models which fail to preserve the essential network architecture of islets. Although the addition of genetic models or isolated receptor measurements would strengthen receptor-specific conclusions, we argue that such approaches alone cannot resolve the physiological complexity of a system where function arises from cell–cell communication and spatiotemporal context.
Indeed, the lack of direct correlation between receptor transcript abundance and functional outcomes has been noted in prior studies, reinforcing the view that function cannot be strictly predicted by molecular presence (Rutter et al., 2024). As articulated in our manuscript, the islet behaves as a sensory collective (Fancher & Mugler, 2017), where emergent patterns— not static cell identity—determine behaviour. This perspective aligns with broader shifts in biology away from strict genetic determinism toward causal emergence and collective agency (Ball, 2023; Levin, 2021).
We therefore believe our study contributes not only new pharmacological insights but also a conceptual reframing of how AVP responses should be interpreted in a complex organ like the pancreas. We have added new data addressing reviewer suggestions—such as glucagon secretion assays, clarifications on the role of forskolin, and an analysis of event timing—that further support our conclusions. We also expanded the discussion on how islet variability is functionally meaningful, not just noise, and explained why β-cell responses to AVP must be interpreted within this probabilistic framework.
We agree that future work should include receptor-specific knockouts and more direct signaling pathway assays, but these would need to be designed with careful consideration of the islet’s dynamic topology and the emergent nature of β-cell roles. In this light, we see our study not as the final word, but as a necessary systems-level foundation for more targeted interventions. We thank the reviewers again for their careful critiques and hope that our response clarifies both the rationale and scope of our work. Our revisions aim to enhance the paper’s clarity while maintaining its commitment to an integrative, physiology-rooted approach.
We thank the reviewers and editors for their thoughtful and constructive assessment of our work. We are especially grateful for their recognition of the study’s technical strengths, including the use of spatio-temporal calcium imaging in intact pancreatic tissue and the strategic development of receptor-selective peptide ligands. We also appreciate their acknowledgement that our study contributes to the understanding of glucose-dependent AVP effects in islet physiology. The reviewers’ concerns regarding variability, receptor specificity, and functional validation helped us further clarify the scope and context of our study.
We respectfully submit that some reservations stem from a reductionist framing that may not fully account for the collective behaviour of islets. As we and others have shown, β-cell function arises from emergent, self-organizing network dynamics, not just from static gene expression or receptor abundance (Jin et al., 2025; Korošak et al., 2021; Postic et al., 2023). In this view, pharmacological heterogeneity across islets is not simply noise or experimental inconsistency, but a signature of dynamic attractor states within the islet network (Stožer et al., 2021). Because an islet functions as a coupled system, most response variability originates from its emergent collective behavior, which eclipses variability in receptor expression or metabolic state.
For this reason, even single-islet receptor quantification or ATP measurements would provide limited explanatory power: it is the state of the network—not absolute receptor levels—that determines whether a perturbation elicits activation or inhibition. As we illustrate in our graphical abstract, a single islet tested repeatedly under identical glucose conditions can yield divergent responses, simply because it occupies different dynamic states. These findings are in line with systems biology and network science approaches, which have revealed that cell function, especially in the β-cell collective, cannot be fully understood through reductionist parameters alone (Gosak et al., 2018; Ball, 2023).
We have included glucagon secretion assays and new analyses to address key reviewer suggestions. Still, we chose not to pursue extensive knockouts or cAMP imaging, as these would require a different experimental scope and could risk disrupting the very dynamics we aim to understand. Likewise, while direct measurements of V1bR or IP3R expression would add molecular detail, they are not definitive without network context. The bell-shaped AVP dose-response curve and its explanation through IP3R inactivation are supported by prior studies; we invoke this mechanism not speculatively, but because it provides the most parsimonious explanation for the glucose-dependent shift in β-cell responsiveness.
We also clarify that our study does not aim to resolve every mechanistic detail, but rather to offer a systems-level insight into how AVP modulates islet dynamics across varying glucose and cAMP contexts. The implications extend beyond AVP pharmacology, suggesting that perturbations to β-cell function must be understood within a probabilistic, state-dependent framework (Fancher & Mugler, 2017). This resonates with emerging concepts in cell physiology that emphasize causal emergence and local agency over static molecular determinism (Levin, 2021; Rutter et al., 2024).
In summary, we see our work as part of a necessary shift in perspective—from linear receptor-function models to context-sensitive dynamic systems. We are grateful for the opportunity to revise our manuscript in response to insightful feedback and hope our clarifications and new data will strengthen its impact for the islet research community.
Public Reviews:
Reviewer #1 (Public review):
Summary:
The authors confirmed earlier findings that AVP influences α and β cells differently, depending on glucose concentrations. At substimulatory glucose levels, AVP combined with forskolin - an activator of cAMP -did not significantly stimulate β cells, although it did activate α cells. Once glucose was raised to stimulatory levels, β cells became active, and α cell activity declined, indicating glucose's suppressive effect on α cells and permissive effect on β cells. Under physiological glucose levels (8-9 mM), forskolin enhanced β-cell calcium oscillations, and AVP further modulated this activity. However, AVP's effect on β cells was variable across islets and did not significantly alter AUC measurements (a combined indicator of oscillation frequency and duration). In α cells, forskolin and AVP led to increased activity even at high glucose levels, suggesting that α cells remain responsive despite expected suppression by insulin and glucose.
Experiments with physiological concentrations of epinephrine suggest that AVP does not operate via Gs-coupled V2 receptors in β cells, as AVP could not counteract epinephrine's inhibitory effects. Instead, epinephrine reduced β cell activity while increasing α cell activity through different G-proteincoupled mechanisms. These results emphasize that AVP can potentiate αcell activation and has a nuanced, context-dependent effect on β cells.
The most robust activation of both α and β cells by AVP occurred within its physiological osmo-regulatory range (~10-100 pM), confirming that AVP exerts bell-shaped concentration-dependent effects on β cells. At low concentrations, AVP increased β cell calcium oscillation frequency and reduced "halfwidths"; high concentrations eventually suppressed β cell activity, mimicking the muscarinic signaling. In α cells, higher AVP concentrations were required for peak activation, which was not blunted by receptor inactivation within physiological ranges.
Attempting to further dissect the role of specific AVP receptors, the authors designed and tested peptide ligands selective for V1b receptors. These included a selective V1b agonist; a V1b agonist with antagonist properties at V1a and oxytocin receptors; and a selective V1a antagonist. In pancreatic slices, these peptides seem to replicate AVP's effects on Ca2+ signaling, although responses were highly variable, with some islets showing increased activity and others no change or suppression. The variability was partly attributed to islet-specific baseline activity, and the authors conclude that AVP and V1b receptor agonists can modulate β cell activity in a statedependent manner, stimulating insulin secretion in quiescent cells and inhibiting it in already active cells.
We applaud the reviewer to capture the essence of work in their introduction.
Strengths:
Overall, the study is technically advanced and provides useful pharmacological tools. However, the conclusions are limited by a lack of direct mechanistic and functional data. Addressing these gaps through a combination of signaling pathway interrogation, functional hormone output, genetic validation, and receptor localization would strengthen the conclusions and reduce the current (interpretive) ambiguity.
Thank you!
Weaknesses:
(1) The study is entirely based on pharmacological tools. Without genetic models, off-target effects or incomplete specificity of the peptides cannot be fully ruled out.
We partially agree with this comment and acknowledge that genetic models would provide a valuable complementary approach to address possible off-target effects or incomplete peptide specificity. However, genetic models also have important limitations, particularly when the aim is to resolve subtle, population-level physiological differences in beta cell activity. We therefore used pharmacological tools at different concentrations to test whether the observed effects were concentration-dependent and consistent with the expected receptor-mediated actions. An advantage of the pancreatic slice preparation is that it preserves much of the native tissue environment and allows pharmacological manipulation within concentration ranges closer to in vivo efficacy, thereby reducing the likelihood of nonspecific effects. To compensate for the lack of genetic models, we now emphasize the collective activity analysis as an additional strength of the study and have clarified this limitation in the revised manuscript.
(2) Despite multiple claims about β cell activation or inhibition, the functional output - insulin secretion - is weakly assessed, and only in limited conditions. This aspect makes it very hard to correlate calcium dynamics with physiological outcomes.
We agree that the functional output needed stronger support and have therefore expanded the hormone secretion experiments. While the effects of AVP and its analogues were tested during a stable plateau phase in the Ca2+ imaging experiments, this phase provides only a narrow dynamic range for insulin release measurements in mouse slices. We therefore added a sequence of stimulations on the same slices, using 8 mM glucose and 500 nM forskolin, with glucose lowered to a non-stimulatory range between different AVP concentrations. These new experiments better define how AVP-dependent changes in Ca2+ dynamics translate into insulin secretion under conditions with a broader secretory dynamic range. The new insulin and glucagon secretion data have now been added to the manuscript as Figure 5, and the text has been revised accordingly.
(3) Insulin and glucagon secretion assays should be provided; the authors should measure hormone release in parallel with Ca2+ imaging, using perifusion assays, especially during AVP ramp and peptide ligand applications.
We added insulin and glucagon secretion assays for AVP ramp to Figure 5.
Additionally, there is no standardization of the metabolic state of islets. The authors should consider measuring islet NAD(P)H autofluorescence or mitochondrial potential (e.g., using TMRE) to control for metabolic variability that may affect responsiveness.
We agree that standardization of the metabolic state of the islets would further strengthen the interpretation of the responsiveness data. We attempted to address this experimentally, but the results were inconclusive and therefore not included in the manuscript. Based on our previous unpublished observations, NAD(P)H levels appear to be significantly higher and less variable in islets within tissue slices than in isolated islets, suggesting that the slice preparation may better preserve the native metabolic state. However, we acknowledge that this remains an important limitation and we now indicate that in the manuscript. Additional experiments will be required to establish a robust and standardized approach, for example by combining NAD(P)H autofluorescence and/or mitochondrial potential measurements with Ca2+ imaging.
(4) There is a high degree of variability in response to AVP and V1b agonists across islets (activation, no effect, inhibition). Surprisingly, the authors do not fully explore the cause of this heterogeneity (whether it is due to receptor expression differences, metabolic state, experimental variability, or other conditions).
This is a well-taken point and has indeed been one of the major bottlenecks in interpreting the results of this study. We agree that the variability in responses to AVP and V1b agonists may reflect several factors, including receptor expression, metabolic state, experimental conditions, and differences in the functional state of individual islets. However, our data also suggest that the beta cell population within an islet should be considered as a dynamic, non-linear system, in which even small differences in initial conditions or collective state can result in qualitatively different outcomes, including activation, no apparent effect, or inhibition. In this framework, the response to AVP is not determined by receptor expression alone, but by the current physiological context of the islet network. This is also why we believe that pharmacological tests are most informative when interpreted within a defined functional state rather than as isolated receptor-specific readouts. As indicated in the graphical abstract, apparently similar islets may occupy different dynamic states and therefore respond differently to the same Gq/PLC/IP3R stimulus. We have now expanded the discussion to make this interpretation more explicit and to acknowledge that receptor expression, metabolic variability, and experimental factors remain possible contributors that will require further targeted studies.
The following text has been added to expand the discussion:
“The heterogeneous responses observed across different islets, where some showed increased activity while others showed no detectable change or inhibition, could intuitively be attributed to variability in V1b receptor expression or signaling capacity among β cells. Such an explanation would be consistent with differences in receptor density, coupling efficiency to Gq proteins, or downstream signaling components such as PLC or IP3 receptors. However, our data suggest that receptor-level variability alone is unlikely to fully explain the observed response spectrum, and that the current functional state of the islet collective must also be considered. The islet behaves as a non-linear dynamic system in which the same molecular perturbation can produce different functional outcomes depending on the current state of the β-cell collective. In such systems, cells or cell populations do not occupy a single deterministic activity state, but rather move within a landscape of possible states, with perturbations shifting the probability distribution of transitions between them. This concept is well established in dynamical systems approaches to biological cell-state transitions, where attractor landscapes, noise, and signaling inputs determine the probability of moving between alternative functional states rather than enforcing a single fixed output.
In this framework, AVP and V1b receptor-selective agonists may reshape the probability landscape of β-cell activity. Depending on the initial metabolic, electrical, and Ca2+-handling state of the islet, the same stimulus may increase oscillation frequency, produce little detectable effect, or shift the system toward reduced activity or functional inactivation. This interpretation is also consistent with studies of pancreatic islet dynamics showing that βcell Ca2+ activity emerges from coupled electrical, metabolic, and network interactions rather than from the properties of individual cells alone. Thus, molecular variability in V1b receptor expression or signaling capacity may contribute to the heterogeneous responses, but it is unlikely to determine them without considering the collective dynamic state of the islet.”
(5) There is no validation of V1b receptor expression at the protein or mRNA level in α or β cells using in situ hybridization, immunohistochemistry, or spatial transcriptomics.
We agree with the reviewer that spatial validation of V1b receptor expression is important for interpreting the cellular targets of AVP signaling in the islet. We have therefore added RNAscope in situ hybridization data to the revised manuscript to assess V1b receptor mRNA expression within the pancreas and islet. These new data show a broader expression pattern of V1b receptor transcripts within the islet than originally assumed, suggesting that AVP signaling may not be restricted to a single endocrine cell population. At the same time, the RNAscope analysis confirms previous reports of higher AVP receptor expression in glucagon-positive alpha cells. We have added these results to Figure 1 and revised the corresponding Results and Discussion sections to clarify that the observed functional responses may reflect both direct effects on beta cells and indirect intra-islet effects mediated through alpha-cell signaling.
(6) AVP effects are described in terms of permissive or antagonistic effects on cAMP (especially in relation to epinephrine), but direct measurements of cAMP in α and β cells are not shown, weakening these conclusions. The authors should use Epac-based cAMP FRET sensors in α and β cells to monitor the interaction between AVP, forskolin, and epinephrine more conclusively.
We agree that direct measurements of cAMP dynamics in alpha and beta cells would provide a more conclusive assessment of the interaction between AVP, forskolin, and epinephrine signaling. We attempted to address this experimentally; however, within the time domain of the Ca2+ oscillations analyzed here, the temporal resolution and robustness of currently available cAMP readouts were not sufficient to resolve these interactions reliably. Even at slower time scales, cAMP sensor signals can be difficult to interpret quantitatively and may be overinterpreted if not tightly linked to the functional readout. We have therefore moderated the wording of the manuscript and now describe the proposed permissive or antagonistic interaction between AVP/V1b and cAMP-dependent signaling as an interpretation supported by the pharmacological Ca2+ response patterns, rather than as a directly demonstrated cAMP mechanism. We now explicitly acknowledge in the limitations that most experiments were performed under cAMP-permissive conditions, which increases sensitivity for detecting AVP-dependent modulation but complicates the separation of direct beta cell effects from intra-islet interactions. Future studies using optimized cell-type-specific Epac-based sensors will be required to resolve this interaction.
(7) Single-islet transcriptomics or proteomics (also to clarify variability) should be provided to analyze receptor expression variability across islets to correlate with response phenotypes (activation vs inhibition). Alternatively, the authors could perform calcium imaging with simultaneous insulin granule tracking or ATP levels to assess islet functional states.
We agree that single-islet transcriptomics, proteomics, or simultaneous metabolic readouts could provide useful complementary information, particularly for describing molecular variability across islets. However, we do not think that differences in receptor expression or ATP levels alone are sufficient to explain the diversity of response phenotypes observed here. Our interpretation is that the beta cell population behaves as a collective dynamic system, in which the same input can lead to different outcomes depending on the current state of the network and its local physiological context. In such a system, AVP/V1b signaling does not necessarily impose a single deterministic response, but changes the probability distribution of accessible states, including activation, inhibition, or no detectable response. Theoretically and partially confirmed by the preliminary data, even the same islet exposed repeatedly under apparently identical conditions could be expected to display different responses if it occupies a different position within this dynamic state space at the time of stimulation. This concept is summarized in the graphical abstract and is central to our interpretation of the pharmacological data. We have therefore clarified in the Discussion that receptor expression, ATP levels, and other molecular parameters may modulate the response landscape, but are unlikely to fully define the observed functional phenotype without considering the collective dynamics of the islet.
Added to Discussion section: “In this framework, AVP and V1b receptorselective agonists may reshape the probability landscape of β-cell activity. Depending on the initial metabolic, electrical, and Ca2+-handling state of the islet, the same stimulus may increase oscillation frequency, produce little detectable effect, or shift the system toward reduced activity or functional inactivation. This interpretation is also consistent with studies of pancreatic islet dynamics showing that β cell Ca2+ activity emerges from coupled electrical, metabolic, and network interactions rather than from the properties of individual cells alone (63). Thus, molecular variability in V1b receptor expression or signaling capacity may contribute to the heterogeneous responses, but it is unlikely to determine them without considering the collective dynamic state of the islet.”
(8) While the study implies AVP acts through V1b receptors on β cells, the signaling downstream (e.g., PLC activation, IP3R isoforms involved) is simply inferred but not directly shown.
We agree that downstream signaling was not directly resolved at the level of PLC activation or specific IP3R isoforms. However, we did not infer Gq/PLC/IP3R involvement solely from AVP pharmacology, but used ACh as an independent Gq-coupled receptor reference stimulus in the same pancreatic slice preparation. With ACh concentration ramps, we could reproduce both activation and inactivation patterns observed with AVP/V1b stimulation, supporting the interpretation that these responses arise from modulation of the Gq-dependent Ca2+ signaling axis.
In addition, in prelilminary expriments we could observe that inhibition of Gq activity with YM254890, as well as interference with IP3R-dependent signaling using Xestospongin C, diminished the response, although not completely. This incomplete suppression is important, because it suggests that beta cell Ca2+ homeostasis and collective islet activity are not controlled by a single linear pathway, but by partially redundant and context-dependent mechanisms. We have therefore revised the manuscript to state more cautiously that our data support the involvement of Gq/PLC/IP3R-dependent signaling, while acknowledging that direct measurements of PLC activity and IP3R isoform-specific contributions remain outside the scope of the present study.
(9) The interpretation that IP3R inactivation (mentioned in the title!) underlies the bell-shaped AVP effect is just hypothetical, without direct measurements. Assays in β (and/or α)-cell-specific V1b KO mice and IP3R KO mice must be provided to support these speculations.
We agree that the involvement of IP3R-dependent signaling should be stated with appropriate caution. However, the concept of IP3R inactivation as a mechanism contributing to bell-shaped Gq-dependent Ca2+ responses is not purely hypothetical, since IP3R inactivation has been directly demonstrated in previous studies and provides a parsimonious explanation for the shift from activation to suppression at higher AVP concentrations. In the present study, this interpretation is further supported by the glucose dependence of the AVP concentration-response relationship, where different stimulatory glucose conditions shift the apparent efficacy peak.
We also agree that cell-specific V1b receptor and IP3R knockout experiments would be valuable future approaches. In this respect, we have obtained preliminary results from a small sample of IP3R triple-knockout mice, which cannot yet be fully included because they are part of an ongoing collaboration. In these experiments, supraphysiological AVP concentrations did not produce the IP3R-like beta-cell response pattern observed in controls, namely reduced halfwidth and increased frequency, whereas alpha cell stimulation was preserved similarly to WT slices.
At the same time, we believe that definitive knockout experiments must be carefully designed, because the beta cell population behaves as a dynamic collective system in which the response to AVP depends on the current functional state of the islet, glucose context, and intercellular coupling. We therefore now present IP3R inactivation as a strongly supported mechanistic interpretation rather than as a directly proven mechanism in this study, and we explicitly acknowledge that cell-specific V1b and IP3R genetic models will be required to fully resolve this pathway.
Reviewer #2 (Public review):
Summary:
In this paper, Drs. Kercmar, Murko, and Bombek make a series of observations related to the role of AVP in pancreatic islets. They use the pancreatic slice preparation that their group is well known for. The observations on the slide physiology are technically impressive. However, I am not convinced by the conclusions of this manuscript for a number of reasons. At the core of my concern is perhaps that this manuscript appears to be motivated to resolve 'controversies' surrounding the actions of AVP on insulin and glucagon secretion. This manuscript adds more observations, but these do not move the field forward in improving or solidifying our mechanistic understanding of AVP actions on islets. A major claim in this manuscript is the beta cell expression of the V1b Receptor for AVP, but the evidence presented in this paper falls short of supporting this claim.
Observations on the activation of calcium in alpha cells via V1b receptor align with prior observations of this effect.
I have focused my main concerns below. I hope the authors will consider these suggestions carefully - please be assured that they were made with the intent to support the authors and increase the impact of this work.
We thank the reviewer for their detailed input and support to increase the impact of our work and our understanding of important cellular processes overall. We have considered their suggestions carefully to further expand the strenghts of our approach and analysis.
Strengths:
The main strength of this paper is the technical sophistication of the approach and the analysis and representation of the calcium traces from alpha and beta cells.
Thank you!
Weaknesses:
(1) The introduction is long and summarizes a substantive body of literature on AVP actions on insulin secretion in vivo. There are a number of possible explanations for these observations that do not directly target islet cells. If the goal is to resolve the mechanistic basis of AVP action on alpha and beta cells, the more limited number of papers that describe direct islet effects is more helpful. There are excellent data that indicate that the actions of AVP are mediated via V1bR on alpha cells and that V1bR is a) not expressed by beta cells and b) does not activate beta cell calcium at all at 10 nM - which is the same concentration used in this paper (Figure 4G) for peak alpha cell Ca2+ activation (see https://doi.org/10.1016/j.cmet.2017.03.017; cited as ref 30 in the current manuscript).
We thank the reviewer for this important comment and agree that the literature on AVP actions in vivo is complex, with several possible sites of action outside the islet. We have therefore revised the Introduction to make the rationale more focused and to better separate systemic effects of AVP from studies addressing direct actions on pancreatic islet cells. At the same time, we chose not to restrict the Introduction only to the alpha cell V1bR literature, because one of the aims of the manuscript is precisely to address why AVP effects on insulin secretion have remained difficult to interpret across experimental contexts.
Our results fully confirm a central aspect of the study cited by the reviewer, namely that V1bR activation robustly stimulates alpha cell Ca2+ activity under non-stimulatory glucose conditions, and that 10 nM AVP does not produce a uniform activation of beta cell Ca2+ activity. In fact, in a substantial fraction of beta cell populations, 10 nM AVP failed to activate oscillations, consistent with the view that alpha cells are the more sensitive and more direct cellular target of AVP/V1bR signaling. However, we do not think that the available transcriptomic evidence is sufficient to categorically exclude V1bR expression or functional relevance in beta cells. Re-analysis of the published dataset, together with more recent datasets and our newly added RNAscope data, supports a higher relative expression of V1bR transcripts in alpha than in beta cells, but does not justify treating beta (or non-alpha) cell expression as absent.
We have therefore revised the manuscript to avoid overstating beta cell V1bR expression as a major isolated claim. Instead, we now present the data as evidence that AVP/V1bR signaling acts most prominently through alpha cells, while beta cell responses emerge in a concentration-, glucose-, and statedependent manner within the intact islet. This interpretation is consistent with the reviewer’s concern that 10 nM AVP preferentially activates alpha cells, but it also accommodates our observation that beta cell collective activity can be modulated under defined pharmacological and metabolic conditions. We believe that this is an important distinction, because the absence of a uniform beta cell Ca2+ activation at one AVP concentration does not exclude beta cell modulation by AVP/V1bR signaling within the intact islet network. The Introduction and Discussion have been revised accordingly to clarify that our study does not simply challenge the alpha cell V1bR model, but expands it by examining how AVP-dependent alpha cell activation, possibly lower beta-cell receptor expression, and collective beta cell dynamics interact in the native pancreatic slice preparation.
(2) We know from bulk RNAseq data on purified alpha, beta, and delta cells from both the Huising and Gribble groups that there is no expression of V2a. I will point you to the data from the Huising lab website published almost a decade ago (http://dx.doi.org/10.1016/j.molmet.2016.04.007) - which is publicly available and can be used to generate figures (https://huisinglab.com/dataghrelin-ucsc/index.html). They indicate the absence of expression of not only AVP2 receptors anywhere in the islet, but also the lack of expression of V1bra, V1brb, and Oxtr in beta cells. Instead of the detailed list of expression of these 4 receptors elsewhere in the body, it would be more directly relevant to set up their pancreatic slice experiments to summarize the known expression in pancreatic islets that is publicly available. It would also have helped ground the efforts that involved the generation of the V1aR agonist and V2R antagonist, which confirm these known AVP/OXT receptor expression patterns.
We thank the reviewer for pointing us more directly to the publicly available islet expression datasets. We agree that the expression of AVP/OXT receptors in purified alpha, beta, and delta cells provides an important reference frame for interpreting our pharmacological data, and we have revised the manuscript to summarize these islet-specific datasets more directly rather than emphasizing receptor expression in other organs. These data support the absence or very low expression of V2 receptors in islet endocrine cells and confirm that V1b receptor expression is substantially enriched in alpha cells compared with beta cells.
At the same time, as outlined in our response above, we do not think that the currently available transcriptomic datasets are sufficient to categorically exclude low-level V1bR transcript expression or functional relevance in beta cells within the intact islet. For this reason, we added independent RNAscope validation to assess V1bR transcripts in the pancreatic slice preparation. These data confirm stronger V1bR expression in glucagon-positive alpha cells, while also showing a broader expression pattern within the islet and pancreas.
We have also revised the rationale for the pharmacological experiments using V1aR- and V2R-directed tools. We now present these experiments not as evidence for unexpected receptor expression, but as functional controls that are consistent with the known AVP/OXT receptor expression patterns in pancreatic islets. This better aligns the manuscript with the existing transcriptomic literature while preserving the main physiological question of the study: how AVP/V1bR-dependent signaling reshapes alpha-cell activity and beta-cell collective dynamics in intact pancreatic tissue.
(3) Importantly, the lack of V1br from beta cells does not invalidate observations that AVP affects calcium in beta cells, but it does indicate that these effects are mediated a) indirectly, downstream of alpha cell V1br or b) via an unknown off-target mechanism (less likely). The different peak efficacies in Figure 4G would also suggest that they are not mediated by the same receptor.
We agree with the reviewer that the absence or very low abundance of V1bR transcripts in beta cells in published transcriptomic datasets would not invalidate the observation that AVP modulates beta-cell Ca2+ activity. It does, however, raise the important question of whether this modulation is mediated indirectly through alpha-cell V1bR activation, through V1bR expression in beta cells that is difficult to resolve transcriptomically, or through another mechanism. To address this more directly, we have now added RNAscope data, which confirm relatively stronger V1bR transcript enrichment in glucagon-positive alpha cells, but also show a broader V1bR transcript signal within the islet and pancreas. Thus, while our data support alpha cells as the dominant V1bR-positive endocrine population, they do not support a strict absence of V1bR-associated signaling capacity in the beta cell compartment.
We also agree that different peak efficacies in alpha and beta cells could be interpreted as evidence for distinct receptors or indirect mechanisms. However, we favor a different interpretation: the apparent efficacy of AVP depends strongly on the physiological state in which the cells are tested. This is particularly evident in beta cells, where the AVP efficacy peak shifts with glucose concentration, suggesting that the beta-cell response is shaped by the metabolic and Ca2+-handling context rather than by receptor occupancy alone. In this framework, the same V1bR/Gq-dependent input can generate different downstream Ca2+ outcomes in alpha and beta cells because the two cell types operate in different dynamic regimes.
We have therefore revised the manuscript to acknowledge this dilemma more explicitly. We now state that beta cell effects of AVP could include indirect alpha cell-dependent components, but given the magnitude and statedependence of the beta cell Ca2+ response it is unlikely to be driven by alpha cell activation. Instead, our preferred interpretation is that AVP/V1bR signaling acts within the intact islet as a context-dependent perturbation of the collective beta cell Ca2+ system, with IP3R-dependent mechanisms being modulated by glucose-dependent changes in beta cell excitability and intracellular Ca2+ handling.
(4) The rationale for the use of forskolin across almost all traces is unclear. It is motivated by a desire to 'study the AVP dependence of both alpha and beta cells at the same time'. As best as I can determine, the design choice to conduct all studies under sustained forskolin stimulation is related to the permissive actions of AVP on hormone secretion in response to cAMPgenerating stimuli. The permissive actions by AVP that are cited are on hormone secretion, which in many cell types requires activation of both calcium and cAMP signaling. Whether the activation of V1br and subsequent calcium response is permitted by cAMP is unclear. I believe the argument the authors are making here is that the activation of beta cell calcium by AVP is permitted by forskolin. i.e., the cAMP stimulated by it in beta cells. However, the design does not account for the elevation of cAMP in alpha cells and subsequent release of glucagon, particularly upon co-stimulation with AVP, which permits glucagon release by activating a calcium response in alpha cells. This glucagon could then activate beta cells. If resolving the mechanism of action is the goal, often less is more. The activation of Gaq-mediated calcium is not cAMP dependent (although the downstream hormone secretion clearly often is). As was shown, AVP does not activate calcium in beta cells in the absence of cAMP. The experiments in Figures 1, 2, and 4 should have been completed in the absence of cAMP first.
We agree with the reviewer that the use of forskolin needs to be explained more clearly, and we have revised the manuscript accordingly. Our rationale was based on the established permissive role of cAMP in AVP-dependent endocrine responses, but we acknowledge that this does not necessarily imply that the upstream V1bR/Gq-mediated Ca2+ response itself is cAMP-dependent. The reviewer is also correct that forskolin elevates cAMP broadly and therefore may affect both alpha and beta cells, including the possibility that AVP-enhanced alpha cell activation and glucagon release secondarily influence beta cell activity.
In fact, our initial experiments were performed without forskolin and revealed an important difficulty: stimulatory glucose alone can increase cAMP levels to a variable extent, as also supported by our previous work on epinephrine signaling, thereby shifting the apparent peak efficacy of AVP stimulation. Thus, forskolin was originally used to reduce this variability and create a more defined cAMP-permissive background in which alpha and beta cell responses could be compared in the same slice. However, we agree that this design works against isolation of beta cell-autonomous AVP effects.
Within a scope of another study we have done an independent series of more focused experiments using GLP-1 receptor stimulation, which preferentially increases cAMP signaling in beta cells compared with the broad cAMP elevation produced by forskolin. We have clarified that the modulation of the AVP-dependent pathway by GLP-1 and related ligands at largely supports beta cell-autonomous AVP effects. It is part of ongoing work and will be reported independently, because a full mechanistic dissection of cAMP–AVP interactions goes far beyond the scope of the present study.
(5) It is unexpected that epinephrine in Figure 2 does not activate the alpha cell calcium? A recent paper from the same group (Sluga et al) shows robust calcium activation in alpha cells in a similar prep by 1 nM epinephrine, which is similar to the dose used here.
We thank the reviewer for pointing this out, but we would like to clarify that epinephrine did significantly activate alpha-cell Ca2+ activity in our experiments, as shown in Fig. 3F. This result is consistent with our previous study by Sluga et al., where low nanomolar epinephrine robustly activated alpha cell Ca2+ signals in the pancreatic slice preparation. The main point of the present comparison was therefore not that epinephrine is inactive in alpha cells, but that AVP produces a substantially stronger and reproducible alpha cell Ca2+ response under comparable experimental conditions. This is also consistent with the data of van der Meulen et al., supporting the view that AVP/V1bR signaling is a particularly potent activator of alpha cell activity. We have revised the text to make this comparison clearer and to avoid the impression that epinephrine failed to activate alpha cells in our preparation.
(6) Figure 8 suggests a pharmacological activation of beta cell V1bR in the low pM range. How do the authors reconcile this comparison with the apparent absence of an effect of AVP stimulation at low pM to low nM doses in beta cells (Figure 4A)? I note that there are changes over time with sustained beta cell stimulation with 8 mM glucose, but these changes are relatively subtle, gradual, and quite likely represent the progression of calcium behaviors that would have occurred under sustained glucose, irrespective of these very low AVP concentrations. I will note that the Kd of the V1bR for AVP is around 1 nM, with tracer displacement starting around 100 pM according to the data in figure 5B, which is hard to reconcile with changes in beta cell calcium by AVP doses that start 10-100-fold lower than this dose at 1 and 10 pM (Figure 8).
We agree that the interpretation of low-pM AVP effects requires caution, particularly when compared with reported V1bR binding affinities. The apparent discrepancy between Fig. 4A and Fig. 8 most likely reflects differences in experimental design, stimulation context, and readout sensitivity. In Fig. 4A, we assessed acute AVP effects under conditions in which beta cell Ca2+ responses are relatively threshold-dependent and where low AVP concentrations produced little or no activation. In contrast, Fig. 8 analyzes prolonged beta cell population dynamics during sustained stimulation with 8 mM glucose, a physiological stimulatory context in which even weak modulatory inputs may become detectable at the level of collective Ca2+ activity.
Importantly, the strongest and statistically significant effect was observed at 100 pM AVP, while lower pM concentrations showed only a trend. We therefore do not interpret the low-pM range as evidence for robust direct pharmacological activation of beta cell V1bR. Rather, these data suggest that AVP may exert permissive or modulatory effects within an already active beta cell network, where glucose-dependent excitability, receptor-effector coupling, and Ca2+ amplification mechanisms can enhance the apparent efficacy of weak inputs. This interpretation is consistent with the known permissive role of AVP in endocrine responses, where AVP may not act as a primary activator alone but can increase the efficacy of other physiological stimuli.
We have also clarified that sustained 8 mM glucose alone does not account for these effects, since Suppl. Fig. 1 shows no comparable time-dependent progression of Ca2+ behavior under sustained glucose stimulation alone. Thus, we now present the low-concentration AVP effects as subtle, contextdependent modulation within the physiological stimulatory range, rather than as evidence for direct beta cell activation at concentrations below the expected receptor affinity range.
Reviewer #3 (Public review):
Summary:
This work aims to better understand the role of arginine vasopressin (AVP) in the control of islet hormone secretion. This builds on previous literature in this area reporting on the actions of AVP to stimulate islet hormones. The gap in literature being addressed by these studies is primarily focused on the glucose-dependency of AVP on both insulin and glucagon secretion. A secondary objective is to explore the role of individual receptors with the use of newly generated peptides and existing tools. The methods include the use of Ca2+ imaging in pancreas slices from mice, with additional outcomes including insulin secretion in some areas. The conclusions presented are that AVP acts through V1b receptors in both alpha- and beta-cells, that this activity occurs in the high cAMP environment, and is glucose dependent.
Strengths:
The area of research is emerging with plenty of room for new contributions. The concept of AVP stimulating islet hormone secretion is important and deserving of further insight. The use of pancreas tissue to image primary cells makes the experiments physiologically relevant. The advancement of novel tools in this area should be helpful to other groups investigating the actions of AVP.
We would like to thank the reviewer for recognizing the potential of our emerging area of research.
Weaknesses:
The conclusions are only modestly supported by the data and lack experimental depth and rigor. The rationale for only conducting studies at high cAMP conditions is not entirely clear and limits the conclusions that can be made. The use of Ca2+ is helpful, but it is a surrogate for hormone secretion. Additional measurements of hormone secretion are needed to enhance the robustness of these conclusions. Consideration of paracrine effects between alpha- and beta-cells is only superficially made and is likely essential in the context of the experimental design. For instance, there is clear literature that alpha-cells secrete several factors that work in paracrine interactions on beta-cells and autocrine actions back on alpha-cells. Conducting these studies in a high cAMP context only completely overlooks these interactions, skewing the interpretations made by the investigators. Finally, the clarity of the experiments and results could be significantly enhanced.
We thank the reviewer for this balanced assessment and for emphasizing several issues that are central to the interpretation of our study. We agree and now explicitly state in the Limitations, that Ca2+ oscillations are a surrogate readout for hormone secretion and that currently used stimulation protocols are not optimized to directly quantify the relationship between Ca2+ dynamics and secretory output. To address this limitation, we have now expanded the functional part of the study by adding complete insulin and glucagon secretion measurements during AVP concentration ramps. These new data provide a stronger functional framework for interpreting the Ca2+ imaging results, while also clarifying that Ca2+ activity and secretion cannot be assumed to correlate linearly under all stimulation protocols.
We have also revised the rationale for the high-cAMP experimental condition. The original aim was to reduce variability arising from glucose-dependent endogenous cAMP signaling and to study alpha and beta cell responses in a common permissive background. However, we agree that broad forskolin stimulation complicates the interpretation of cell-autonomous versus paracrine mechanisms. Independent experiments within a scope of another study demonstrate that using GLP-1 co-stimulation, which provides a more beta-cell-oriented cAMP-permissive condition and supports the interpretation that AVP can modulate beta-cell collective activity in a manner that is not solely secondary to alpha-cell activation.
We fully agree that paracrine interactions within the islet are physiologically important and must be considered, particularly in intact pancreatic slices. Nevertheless, the rapid onset of the AVP effects observed in beta cell Ca2+ activity argues against a mechanism mediated predominantly by slower indirect paracrine loops. High AVP concentrations, as shown in Fig. 5, significantly shorten the intervals between Ca2+ events in both alpha and beta cells, but that the activity of the two cell populations remains largely noncoordinated. This temporal dissociation does not exclude paracrine modulation altogether, but it argues against a simple alpha-cell-driven explanation for the beta-cell response.
We have revised the manuscript to state these points more clearly and to moderate conclusions where the data support modulation rather than definitive cell-autonomous receptor action. We believe that the added secretion experiments and alpha/beta event-timing analysis substantially strengthen the physiological interpretation of the study, and we thank the reviewer for raising these issues.
Recommendations for the authors:
Reviewer #2 (Recommendations for the authors):
(1) The paragraph discussing the benefits of slice physiology over islets is not reflective of how most - if not all of your colleagues who do islet experiments conduct these. Many labs have reported for years high-quality GSIS experiments, synchronous calcium responses, and a plethora of studies detailing the mechanism of hormone and neurotransmitter actions using islet models, and have done so well. Slice physiology is a unique and helpful model that can have advantages over other models. This particular reviewer uses both models in their lab and each has benefits and - inevitably - drawbacks. Many of the possible drawbacks cited for islet studies apply equally to slices, including the possibility of altered gene expression, lack of innervation, and circulation. Added drawbacks are the exposure to higher levels of pancreatic enzymes from the slice, which require co-culture with enzyme inhibitors.
We agree with the reviewer and have revised these limitations accordingly. Our intention was not to imply that isolated islet preparations are generally inferior, since they have provided a highly productive and rigorous experimental platform for GSIS, synchronized Ca2+ dynamics, and mechanistic studies of hormonal and neurotransmitter regulation with standardized protocols with all their positive and negative sides. We modified the presentation of pancreatic slices as a complementary model with specific advantages, particularly preservation of local tissue architecture, while also acknowledging their limitations. The revised text therefore avoids a comparative hierarchy between slices and isolated islets and instead emphasizes that both models have distinct strengths and drawbacks depending on the experimental question.
(2) If you want to demonstrate direct actions on beta cells, deconstructing the islet would be a better way to go. Less complicated, not more. Dissociated beta cells, instead of slices, were used just to prove or disprove the hypothesis of direct beta cell effects of AVP.
We agree that dissociated beta cells can be a useful reductionist model to test whether AVP is capable of acting directly on individual beta cells. However, this approach would also remove the collective beta-cell activity that is central to the physiological question addressed in the present study. Since our data indicate that AVP effects emerge within the intact islet as rapid and extensive changes in coordinated Ca2+ dynamics, dissociation would not necessarily provide a more informative model for understanding these responses. The fast onset and magnitude of the beta cell response argue against a predominantly indirect non-autonomous mechanism, and this interpretation is further supported by the GLP-1 co-stimulation experiments, which are more consistent with beta cell-autonomous modulation. We have therefore clarified in the revised manuscript that dissociated-cell experiments would be valuable for a narrowly defined receptor-cell autonomy question, but would not resolve the collective islet dynamics that are the focus of this work.
(3) If you want to sustain the claim of beta cell expression of V1br, you would have to demonstrate this far more directly by staining (if appropriate antibodies exist), by beta cell-specific deletion of V1br, or by highly selective, well-validated pharmacology. This should include a demonstration of Gaqdependence in isolated beta cells.
We have added RNAscope in situ hybridization data to the revised manuscript to assess V1b receptor mRNA expression within the pancreas and islet. These new data show a broader expression pattern of V1b receptor transcripts within the islet than originally assumed, suggesting that AVP signaling may not be restricted to a single endocrine cell population. At the same time, the RNAscope analysis confirms previous reports of higher AVP receptor expression in glucagon-positive alpha cells. We have added these results to Figure 1 and revised the corresponding Results and Discussion sections to clarify that the observed functional responses may reflect both direct effects on beta cells and indirect intra-islet effects mediated through alpha-cell signaling.
Minor
(1) O'Carroll et al. should be cited in the context of islet permissive actions of AVP/cAMP. PMID: 18434353, although that paper offers no evidence that the AVP-dependent potentiation of insulin release is mediated directly by beta cells. It does confirm dependence on PKC.
We agree and have added O’Carroll et al. in the revised manuscript in the context of AVP/cAMP-dependent permissive actions on islet hormone secretion. We therefore use it as support for the broader concept of AVP-dependent amplification of secretion in a permissive signaling context, rather than as direct evidence for beta-cell-autonomous V1bR signaling.
(2) Figure 2 E-H, glucose concentration mislabeled.
We thank the reviewer for pointing this out. The glucose concentration label has been clarified: panels E–H show pooled data from separate experiments performed at 8 mM glucose, whereas panel D shows a representative experiment performed at 9 mM glucose.
(3) The insulin secretion in 4E is difficult to interpret without a low-glucose control. If this is hard to do in a slice preparation, a separate static islet secretion experiment would help here. The possibility that the inhibition of insulin secretion traces back to the activation of delta cells by AVP could be considered - I struggle to come up with a plausible mechanistic explanation why AVP (which activates calcium in alpha and in beta cells in the presence of 8 mM G plus forskolin according to your data) would inhibit insulin secretion.
We agree that the original insulin secretion experiment was difficult to interpret without a clearer low-glucose reference condition. To address this, we have added new insulin release experiments in which glucose was lowered to a non-stimulatory range between AVP concentrations, followed by sequential stimulation with 8 mM glucose and 500 nM forskolin in the same slices. These new data provide a broader dynamic range for assessing insulin secretion and allow a more direct comparison between AVP-dependent Ca2+ modulation and secretory output.
We also agree that AVP-dependent inhibition of insulin secretion requires careful interpretation. One possible explanation is not simply activation of delta cells, but a failure of the beta cell collective to maintain coordinated activity at very high AVP concentrations. In the Ca2+ imaging data, high AVP concentrations increase activity in many beta cells, but numerous cells within the islet fail to keep pace with the collective oscillatory rhythm, leading to fragmented and less synchronized population activity. Thus, despite increased frequency of Ca2+ oscillations in the islet, the integrated beta cell output may become less efficient for insulin secretion. We have added this interpretation to the revised manuscript and now discuss delta cell activation as a possible contributing mechanism, but not as the primary explanation supported by our current data.
