Insulin and Incretin Receptor Agonists Reciprocally Alter Their Blood–Brain Barrier Permeabilities

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper investigates the interaction between insulin and GLP-1 on transport across the BBB. They find that an insulin receptor antagonist did not affect the BBB influx of radiolabeled dulaglutide, a GLP-1 receptor agonist. However, insulin increased the transport of dulaglutide across the BBB. The uptake was fastest in the olfactory bulb, and slower in other areas. Insulin uptake was not affected by dulaglutide, but insulin receptor affinity was reduced. Semaglutide reduced the uptake of insulin into the brain, too, but not of leptin.

The study is of interest, as there is more research ongoing into the roles of insulin and of GLP-1 in the brain nowadays. Key brain areas in the cortex and limbic system express GLP-1 receptors and have shown effects on mental states/mood disorders, drug addiction, appetite and neurodegeneration. The study has been conducted at a high level. I only have a few comments to make.

 Why did the authors choose dulaglutide as a GLP-1 receptor agonist? It is a very large molecule and is designed to stay in the blood to treat diabetes. Dulagutide has a half -life in the blood of 90 hours, and the MW is 60,000. It is surprising that such a molecule would enter the brain at meaningful rates after only 6 min. If it stays in the blood for a long time, how can it enter the brain at such rapid rates? The drug cannot be in two places at the same time. Are the authors sure that they are working with the dull peptide and not some fragment? Exenatide is a much smaller peptide and has a short half-life in the blood. It can cross the BBB at a reasonable rate. Would that not be a better choice?

Author Response

The paper investigates the interaction between insulin and GLP-1 on transport across the BBB. They find that an insulin receptor antagonist did not affect the BBB influx of radiolabeled dulaglutide, a GLP-1 receptor agonist. However, insulin increased the transport of dulaglutide across the BBB. The uptake was fastest in the olfactory bulb, and slower in other areas. Insulin uptake was not affected by dulaglutide, but insulin receptor affinity was reduced. Semaglutide reduced the uptake of insulin into the brain, too, but not of leptin.

The study is of interest, as there is more research ongoing into the roles of insulin and of GLP-1 in the brain nowadays. Key brain areas in the cortex and limbic system express GLP-1 receptors and have shown effects on mental states/mood disorders, drug addiction, appetite and neurodegeneration. The study has been conducted at a high level. I only have a few comments to make.

 Why did the authors choose dulaglutide as a GLP-1 receptor agonist? It is a very large molecule and is designed to stay in the blood to treat diabetes. Dulagutide has a half -life in the blood of 90 hours, and the MW is 60,000. It is surprising that such a molecule would enter the brain at meaningful rates after only 6 min. If it stays in the blood for a long time, how can it enter the brain at such rapid rates? The drug cannot be in two places at the same time. Are the authors sure that they are working with the dull peptide and not some fragment? Exenatide is a much smaller peptide and has a short half-life in the blood. It can cross the BBB at a reasonable rate. Would that not be a better choice?

Author Response: Dulaglutide was selected as the “acute” GLP-1RA as we previously identified it exhibited a faster transport rate than exenatide (PMID: 38095516, Figure 3). For our acute studies, we wanted an IRA that exhibited a fast rate of BBB transport. During the course of these studies, we identified the dulaglutide peptide we were working with was indeed a fragment of the full-length molecule, which we have carefully referred to as the bioactive fragment (BAF) throughout our manuscript, with a molecular weight of 3,315 Da. While we have been sure to refer to this peptide as the “BAF” throughout the manuscript, we have now included the molecular weight of this peptide in the Methods section, line 405.

Reviewer 2 Report

Comments and Suggestions for Authors

1- Clearly state in the abstract and discussion that major findings (chronic effects on insulin transport) are female-selective; include sex-stratified data more prominently throughout.

2- Report the exact "n" for each group/region in all tables and figures. Add statistical annotations (p values) directly on Figure 2 and Figure 5 for clarity.

3- Expand discussion on why certain regions (e.g., olfactory bulb, pons) show faster IRA transport and how this relates to known receptor distribution or lipid solubility.

4- Acknowledge the relatively short 2-week treatment and absence of obesity/disease models; discuss relevance to long-term clinical IRA use.

5- Clarify why olfactory bulb data were too variable for analysis and whether this affects interpretation of whole-brain findings.

6- Improve Figure 2 by using consistent scales or separating panels; ensure all linear regression lines include r and p values in legends.

7- Discuss potential mechanisms (e.g., receptor desensitization, transporter modulation, or endothelial signaling) for insulin-enhanced IRA transport and chronic IRA-reduced insulin transport.

Author Response

1- Clearly state in the abstract and discussion that major findings (chronic effects on insulin transport) are female-selective; include sex-stratified data more prominently throughout.

Author Response: We have now stated specifically that the chronic effects of semaglutide on insulin transport are female-selective. We have read through the manuscript and described the sex-stratified data more clearly and indicated when males and females were combined in the analysis. Examples of edited language include “female-selective” and “male and female mice combined”.

2- Report the exact "n" for each group/region in all tables and figures. Add statistical annotations (p values) directly on Figure 2 and Figure 5 for clarity.

Author Response: We apologize for this oversight as we showed each individual data point within the figure. For clarity, we have now included the “n” for each group within the Figure Legend or respective Table.  The statistical annotations are already present in Figure 2, but we have included the definition of each annotation in Figure 2 legend. Figure 5 now has statistical annotation added to panel C on the y-axis for the significant difference between the Vi (y-intercepts). To avoid confusion, we do not depict the statistical annotation for the significance of the linear regression (i.e., p Ki) on the graph and instead, reserve the statistical annotation for data where the slopes of the two lines (Ki) are significantly different.

3- Expand discussion on why certain regions (e.g., olfactory bulb, pons) show faster IRA transport and how this relates to known receptor distribution or lipid solubility.

Author Response: The olfactory bulb is a metabolic regulator of energy homeostasis, modulating odor-driven food intake. GLP-1R signaling in the olfactory bulb contributes to food intake and metabolism. The pons/medulla (hindbrain area) is involved in visceral feedback for suppressing food intake and reducing appetite. Both the olfactory bulb and the pons/medulla exhibit high expression of the GLP-1R and the GIPR. Please see lines 282-286.

4- Acknowledge the relatively short 2-week treatment and absence of obesity/disease models; discuss relevance to long-term clinical IRA use.

Author Response: We thank this reviewer for these key comments. We selected the 2-week treatment period as this is a time when weight loss plateaus in mice (PMID: 32213703, Gabery et al 2020 JCI, Figure 1A). We utilized lean mice in this study as there is accumulating evidence about the beneficial effects of IRAs in multiple diseases (substance use disorder, mild cognitive impairment) co-morbid with obesity. However, we anticipate next clinical steps will be to investigate IRA use in states without obesity. Studying lean mice provides us with information about basal activities of IRAs independent of obesity. We have now included this explanation in the Methods, lines 424 and 428-431.

5- Clarify why olfactory bulb data were too variable for analysis and whether this affects interpretation of whole-brain findings.

Author Response: The mouse olfactory bulb is approximately 15 mg compared to 500 mg for whole brain. We relate the amount of radioactivity (CPM) detected in each sample with the weight of the tissue. Because of the small sample weight, minor differences in the level of radioactivity detected contribute to the variability of the olfactory bulb data, which is not as apparent in the whole brain data due to the larger sample weight. We are confident in all data with significant linear regressions, particularly in the whole brain, and this significance adds to our interpretation.

6- Improve Figure 2 by using consistent scales or separating panels; ensure all linear regression lines include r and p values in legends.

Author Response: We are unsure what this comment is in reference to as Figure 2 does not have multiple panels or linear regression lines. All r and p values are reported in the respective Tables. The Figure legend and Table legend call attention to the respective Table or Figure that the data relates to (for example in the Figure 2 legend, it is stated “Full linear regression data and sample size are listed in Table 2”. In the Table 2 legend, it is stated “r is the correlation coefficient for the regression lines between time and brain region/serum ratios shown in Figure 2”.). We have reviewed all figures to ensure they are consistent throughout.

7- Discuss potential mechanisms (e.g., receptor desensitization, transporter modulation, or endothelial signaling) for insulin-enhanced IRA transport and chronic IRA-reduced insulin transport.

Author Response: We have included additional discussion about the potential mechanisms that may be responsible for the changes observed in our present work. Lines 317-321 now state: “Although prior work has focused on the changes in exendin-4 transport into the hypothalamus [29], our work extends these findings to other brain regions and supports changes in energy needs alter IRA BBB transport. Whether these changes are due to receptor desensitization, transporter modulation, or endothelial signaling remains to be determined.” Additionally, lines 370-373 state: “We have shown activation of insulin signaling downregulates insulin BBB transport [18]. Therefore, it is possible that signaling changes within the brain are driving the reductions in insulin BBB transport.”

Reviewer 3 Report

Comments and Suggestions for Authors

Interesting study; The authors report that semaglutide did not cause weight loss in mice. This is strange because semaglutide (Ozempic/Vigovy) is known for its weight loss properties. However, the authors explain that they used lean mice. Please discuss this in more depth in the Discussion section.

I recommend adding a graphic abstract to this manuscript to increase readability and citations.

Given that the authors claim that insulin receptor inhibition (S961) in a previous study [18] slowed insulin transport, but elsewhere in the same text it is reported to have enhanced transport (line 105), is this a typographical inconsistency or is there a difference in the experimental models? 

The authors used the word Slowed on line 100 and enhanced on line 105, which should be clarified.

The authors observed that insulin increased the Ki of dulaglutide. Is this effect due to an increase in systemic blood flow by insulin or a specific mechanism in the endothelial cells of the blood-brain barrier? Please discuss this in more depth in the Discussion section. Did the authors measure regional cerebral blood flow (rCBF) to rule out hemodynamic effects?” If not, please mention this as a suggestion at the end of the article.

How is semaglutide beneficial in animal models of Alzheimer’s when you have shown that it reduces insulin transport (which is critical for brain health) in females? Please mention on this in the discussion section.

The authors initially thought that the reason why insulin entered the female brain less was because there was too much insulin in the blood of female mice, (saturating the receptors). But the results of Figure 4-F showed that the blood insulin of the females actually decreased.  When blood insulin decreases,  transport to the brain should be "increased" (because the receptors are empty), but here it decreased. This is a point which need more clarification. Please discuss in more depth in the discussion section.

Author Response

Interesting study; The authors report that semaglutide did not cause weight loss in mice. This is strange because semaglutide (Ozempic/Vigovy) is known for its weight loss properties. However, the authors explain that they used lean mice. Please discuss this in more depth in the Discussion section.

Author Response: We have clarified the use of lean mice throughout our manuscript, including in the Abstract (line 17). In our response to Reviewer 2, point 4, we utilized lean mice in this study as there is accumulating evidence about the beneficial effects of IRAs in multiple diseases (substance use disorder, mild cognitive impairment) co-morbid with obesity. However, we anticipate next clinical steps will be to investigate IRA use in states without obesity. Studying lean mice provides us with information about basal activities of IRAs independent of obesity. We have now included this explanation in the Methods, lines 416 and 420-423.

I recommend adding a graphic abstract to this manuscript to increase readability and citations.

Author Response: We thank the reviewer for this suggestion. We have generated a new graphic abstract in Biorender that summarizes our main findings to increase readability and citations.

Given that the authors claim that insulin receptor inhibition (S961) in a previous study [18] slowed insulin transport, but elsewhere in the same text it is reported to have enhanced transport (line 105), is this a typographical inconsistency or is there a difference in the experimental models? 

The authors used the word Slowed on line 100 and enhanced on line 105, which should be clarified.

Author Response: We apologize for this error and confusion. Thank you for drawing this discrepancy to our attention. Prior work [18] showed brain insulin receptor inhibition with ICV S961 slowed insulin BBB transport. We have corrected this in the manuscript (line 105).

The authors observed that insulin increased the Ki of dulaglutide. Is this effect due to an increase in systemic blood flow by insulin or a specific mechanism in the endothelial cells of the blood-brain barrier? Please discuss this in more depth in the Discussion section. Did the authors measure regional cerebral blood flow (rCBF) to rule out hemodynamic effects?” If not, please mention this as a suggestion at the end of the article.

Author Response: This is an excellent point of discussion. Regional cerebral blood flow (rCBF) was not measured in this study. The effects of insulin on BBB transport rates of ¹²⁵I-dulaglutide (BAF) were measured within a short time period 0.5-5 min. Prior studies indicate intravenous insulin does not alter CBF (PMID: 39584591). Therefore, we do not believe changes in CBF are driving the increased transport rates for dulaglutide. We have now included this in the Discussion (lines 322-326): “There is evidence indicating IRAs cross the BBB through adsorptive transcytosis [17]. Whether changes in cerebral blood flow can impact adsorptive transcytosis is unclear. Nevertheless, it is unlikely cerebral blood flow is involved in increasing dulaglutide (BAF) BBB transport in our system as intravenous insulin has previously been shown to not alter cerebral blood flow based on prior work [34].

How is semaglutide beneficial in animal models of Alzheimer’s when you have shown that it reduces insulin transport (which is critical for brain health) in females? Please mention on this in the discussion section.

Author Response: We agree with this comment that our findings were opposite of what we hypothesized. We have now included a greater discussion about the regulation of insulin BBB transport (lines 369-374):  “The reductions in insulin BBB transport following semaglutide treatment were opposite of what we hypothesized. While reductions in insulin BBB transport in disease states are often described as negative, the full picture remains to be elucidated. Insulin BBB transport is clearly finely tuned and is regulated by multiple pathways [40]. Determining the molecular mediators of insulin BBB transport will aid in understanding how insulin BBB transport is impacted by various physiological states.”

The authors initially thought that the reason why insulin entered the female brain less was because there was too much insulin in the blood of female mice, (saturating the receptors). But the results of Figure 4-F showed that the blood insulin of the females actually decreased.  When blood insulin decreases,  transport to the brain should be "increased" (because the receptors are empty), but here it decreased. This is a point which need more clarification. Please discuss in more depth in the discussion section.

Author Response: The reviewer is correct that reductions in circulating insulin would enhance the transport rate of radiolabeled insulin due to less competition. Since insulin levels in females were significantly reduced, we wanted to make sure these changes, and other changes in circulatory factors, were not contributing to our observed insulin transport rates. The in situ studies confirmed our in vivo studies. We have now included this discussion in lines 362-369: “Insulin BBB transport is affected by multiple circulating factors including triglycerides and even insulin itself [40]. Chronic semaglutide clearly affected circulating factors of some hormones measured in this study, decreasing insulin and GLP-1 levels in lean female mice. While a reduction in insulin would actually increase transport of the insulin radiotracer due to lack of competition, we saw the opposite, where insulin BBB transport was reduced. Our in situ studies supported this semaglutide reduction of insulin BBB transport, indicating something other than a change in circulating factor is responsible.”