Review Reports - Semaglutide Prevents Aortic Rupture and Dissection in the Angiotensin II Mouse Model

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The present study reports the promising results on semagluatide mediate vascular protection. The findings are novel and acceptable for the publication. The below issues to be addressed by authors.

The introduction is missing information o ApoE role in the introduction. Since authors have used ApOE knock out mice, it must be explained for general audience to understand the ApoE role.
Figure 2B. Increase the arrows' thickness for better visibility.
Line 325-330: Add references
Please mention the limitations and future scope of the present study.
Compare your findings with other GLP1 agonists.

Author Response

The present study reports the promising results on semagluatide mediate vascular protection. The findings are novel and acceptable for the publication. The below issues to be addressed by authors.

The introduction is missing information o ApoE role in the introduction. Since authors have used ApOE knock out mice, it must be explained for general audience to understand the ApoE role.

Reply: We agree, this was now added to the introduction, page 2, row 63-64.

Figure 2B. Increase the arrows' thickness for better visibility.

Reply: We agree, this was now added to the figure.

Line 325-330: Add references

Reply: This is our own data we discuss.

Please mention the limitations and future scope of the present study.

Reply: We mention limitations in the discussion at page 12, row 356-361. We have now also added that the dose chosen could affect the results and comparison to liraglutide, page 12, row 361-364. We also added that sub-group analysis makes the groups small for some analysis and may be underpowered, page 12, row 365-367. At page 12, row 353-354 we also discuss future experiments needed to better understand the mechanisms.

Compare your findings with other GLP1 agonists.

Reply: To our knowledge, the only other GLP-1 agonist that have been studied in relation to AAA is Liraglutide, which we discuss our data to. We have clarified our discussion so its more clear that it is to liraglutide we compare our results. Page 11, row, 329.

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript investigates the protective effect of semaglutide, a GLP-1 receptor agonist, against abdominal aortic aneurysm (AAA) rupture and dissection in angiotensin II-infused ApoE-deficient mice. The study provides novel evidence that prophylactic semaglutide administration significantly reduces early mortality from aortic rupture, likely through maintenance of aortic wall collagen content. The topic is clinically highly relevant given the lack of pharmacological therapies for AAA, and the findings have important translational implications. While the study has significant strengths, several issues need to be addressed to improve the quality and clarity of the manuscript.

The sample sizes across different experimental arms are highly variable (e.g., n=4 in saline groups, n=22 in semaglutide prophylactic group). The power calculation described in the methods specifies an ability to detect 50% changes with 80% power, but it is unclear if this applies to all experimental groups, particularly the smaller subgroups.Please provide explicit power calculation details for each experimental arm, including justification for the varying sample sizes. Consider supplementing with post-hoc power analysis to confirm that negative results (e.g., lack of effect on late AAA progression) are not due to underpowering.
The conclusion that semaglutide maintains collagen content through increased TIMP2 expression is not fully supported by the data. While TIMP2 mRNA is significantly increased in the semaglutide group compared to saline, there is no statistically significant difference in TIMP2 expression between the AngII+vehicle and AngII+semaglutide groups. Additionally, no data on MMP or TIMP protein activity is provided.
The rescue trial shows no significant effect of semaglutide on survival or AAA progression when initiated 7 days after disease induction. However, the high early rupture rate (46% in vehicle group) before treatment initiation may have selected for a more resilient mouse population that is less responsive to intervention.Discuss the potential survivor bias in the rescue trial cohort as a possible explanation for the negative result.Consider modifying the conclusion to clarify that semaglutide's protective effect is limited to the early initiation phase, rather than implying it has no effect on established disease without this important context.
The metabolite identifications from the MALDI-MSI experiments are based solely on accurate mass matching without confirmatory MS/MS for most features, including the galactosylceramide (d18:1/16:0) finding. The authors appropriately note this limitation, but the discussion overstates the causal role of these metabolites in the protective effect.Clearly qualify all metabolite identifications as "putative" or "tentative" unless MS/MS confirmation is available.Tone down the causal interpretation of the metabolomics findings, framing them as hypothesis-generating rather than confirmed mechanisms.
The abstract states that semaglutide "drastically reduced mortality by dissection and rupture" in the prophylactic trial, but does not mention the lack of effect on late AAA progression. This may give readers an incomplete impression of the study findings.Add a brief statement in the abstract noting that semaglutide did not significantly reduce AAA diameter or incidence in surviving animals at 28 days.

Author Response

The sample sizes across different experimental arms are highly variable (e.g., n=4 in saline groups, n=22 in semaglutide prophylactic group). The power calculation described in the methods specifies an ability to detect 50% changes with 80% power, but it is unclear if this applies to all experimental groups, particularly the smaller subgroups. Please provide explicit power calculation details for each experimental arm, including justification for the varying sample sizes. Consider supplementing with post-hoc power analysis to confirm that negative results (e.g., lack of effect on late AAA progression) are not due to underpowering.

Reply: We thank the reviewer for pointing this important detail out and we agree that this can be more clear. Our main readout was difference in aneurysm development between angII and angII+semaglutide treated animals with an assumed difference of at least 50%, and the power calculation was based on this. The high number of animals in the semaglutide and angII vehicle prophylactic group is due to the experience from the rescue trial where 46% of the mice died. The low number of animals in the saline group is since this groups is included as a baseline comparison only where aneurysms are not developed and the aortic diameter is very homogenous without any large variation in size. The aortas of these animals don’t rupture and we don’t loose any animals in this group due to death. The important comparison is between the angII vehicle and angII + semaglutide. We have added to the statistics page 5, row 184 that the power calculation refers to aortic diameter. We further add to the limitation that sub-group analysis have lower number of animals and may be due to underpowering. Page 12, row, 365-366.

2. The conclusion that semaglutide maintains collagen content through increased TIMP2 expression is not fully supported by the data. While TIMP2 mRNA is significantly increased in the semaglutide group compared to saline, there is no statistically significant difference in TIMP2 expression between the AngII+vehicle and AngII+semaglutide groups. Additionally, no data on MMP or TIMP protein activity is provided.

Reply: We don’t conclude that semaglutide maintain collagen content through increased TIMP2 expression. We wanted to determine if the increased collagen content could be a result of altered increase in the production of collagens or decreased degradation. We added to the result section, page 8, row 273 that this is compared to non-induced animals. To clear out for the reader that it is mRNA we have analysed, we have specified this in the result section, page 9, rows 264-275. In the discussion we write that the increase of TIMP2 in the semaglutide treated animals is compared to the non-induced controls, page 11, row 350-351. We also discuss that activity assays are needed to fully determine the mechanisms of semaglutide on collagen production and degradation, page 12, row 356.

3. The rescue trial shows no significant effect of semaglutide on survival or AAA progression when initiated 7 days after disease induction. However, the high early rupture rate (46% in vehicle group) before treatment initiation may have selected for a more resilient mouse population that is less responsive to intervention. Discuss the potential survivor bias in the rescue trial cohort as a possible explanation for the negative result. Consider modifying the conclusion to clarify that semaglutide's protective effect is limited to the early initiation phase, rather than implying it has no effect on established disease without this important context.

Reply: This is an important point and was now added to the discussion, page 11, row 323-325.

4. The metabolite identifications from the MALDI-MSI experiments are based solely on accurate mass matching without confirmatory MS/MS for most features, including the galactosylceramide (d18:1/16:0) finding. The authors appropriately note this limitation, but the discussion overstates the causal role of these metabolites in the protective effect. Clearly qualify all metabolite identifications as "putative" or "tentative" unless MS/MS confirmation is available. Tone down the causal interpretation of the metabolomics findings, framing them as hypothesis-generating rather than confirmed mechanisms.

Reply: This is a reasonable comment and have been added to the figure legends and discussion, page 12, row 376, and page 12, rows 390-392.

Figure 5A) Glutathione (GSH) [M-H]- at m/z 306.0768 was the only candidate metabolite obtained using the met-ID software [Bjärterot et al., 2025]. The identity was further confirmed from specific MSMS-fragmentation.

Figure 5B) Glutathione (GSSG) [M-H]- at m/z 611.1440 was potentially identified based on accurate mass (+0.7 ppm) together with comparison of the distribution of GSH.

Figure 5C) Cholesterol sulphate [M-H]- at m/z 465.3049 was the only candidate metabolite obtained using the met-ID software [Bjärterot et al., 2025].

Figure 5D) Sphingomyelin (33:1) [M-H]- at m/z 687.5452 was the only candidate metabolite obtained using the met-ID software [Bjärterot et al., 2025].

Figure 6) Galactosylceramide (d18:1/16:0) [M+Cl]- at m/z 734.5348 was the only candidate metabolite obtained using the met-ID software [Bjärterot et al., 2025].

5. The abstract states that semaglutide "drastically reduced mortality by dissection and rupture" in the prophylactic trial, but does not mention the lack of effect on late AAA progression. This may give readers an incomplete impression of the study findings. Add a brief statement in the abstract noting that semaglutide did not significantly reduce AAA diameter or incidence in surviving animals at 28 days.

Reply: We agree to this comment that clarifies the abstract and have changed this accordingly, row 26-27.

Reviewer 3 Report

Comments and Suggestions for Authors

The paper by Ramilo et al. demonstrates that semaglutide treatment in the angiotensin II induced AAA model in ApoE KO mice protected against AAA rupture when treatment was initiated at the time of AAA induction. The underlying mechanism was described to increase aortic wall thickness on day 7 after AAA induction, with a trend toward elevated collagen content, which is in line with other publications. Overall, the paper is interesting but also contains several points both in design and analyses that need to be addressed.

For the methods:

Power calculations were performed to identify the number of animals needed to have enough statistical power. Please provide this number. Was it based on aneurysm diameter?

You have access to Vevo11000 high frequency ultrasound why was mice only scanned at endpoint of experiments? Especially for the recue trial, it would have made more sense since the mice are housed mixed to treatment, that the groups were defined based on abdominal aortic diameter at day 7 prior to initiation of semaglutide treatment, as AAA in this model variates a lot. Alos in line that most of the mice died before treatment initiation

Staining for GLP1R in aortic tissue is difficult. There is no mentioning of positive or negative control were used for specificity of the staining. At least negative isotypic control should be included. Is it possible to improve micrograph quality as positive staining in figure 1 is hard to detect.

The numbers in table 1 do not add up. Example: of 22 mice 2 died of AAA rupture leaving 20 lefts. Though only 18 is reported to survive whereof 5 out of 13 developed and AAA.could you please clarify this.

The nicely performed Mass spectrometry imaging is performed on the samples collected on day 7 or on day 28 from the prophylactic trial?

Please describe how you measured wall thickness in the histological samples. If you have a more representative micrograph of the Ang II + vehicle treated group, I recommend using another Ang II vehicle micrograph as the quality is poor in fig 4 C

For statistical analyses of AAA growth and RNA levels. Why do you mainly compare saline control with AngII semaglutide, rather than compare Ang II controls and AngII treated with semaglutide, that to me would be more relevant.

In the discussion, page 11 line 326 you write “at mRNA levels of col1A1 and Col3A1.. were increased both I the vehicle controls and semaglutide treated group”. Data in table 2 does not support this.

Minor comments.

For the description of the prophylactic trail where semaglutide was initiated on day 7 continued to day 28 and not for 28 days.

Author Response

1. For the methods:

Power calculations were performed to identify the number of animals needed to have enough statistical power. Please provide this number. Was it based on aneurysm diameter?

Reply: Yes, this was based on aneurysm diameter and have now been clarified in the statistic section, page 5, row 185.

2. You have access to Vevo11000 high frequency ultrasound why was mice only scanned at endpoint of experiments? Especially for the recue trial, it would have made more sense since the mice are housed mixed to treatment, that the groups were defined based on abdominal aortic diameter at day 7 prior to initiation of semaglutide treatment, as AAA in this model variates a lot. Also in line that most of the mice died before treatment initiation.

Reply: This is a valid comment. Our experience with the model is that the mice are affected by the surgery and angII during the first week and drops up to 20% in weight before they start to gain weight again. We simply don’t want to stress them more by sleeping them down and perform a 30 minute experiment on them. This is also something that can complicate the ethical approval when we apply for it in Sweden where it has become more difficult over time.

3. Staining for GLP1R in aortic tissue is difficult. There is no mentioning of positive or negative control were used for specificity of the staining. At least negative isotypic control should be included. Is it possible to improve micrograph quality as positive staining in figure 1 is hard to detect.

Reply: Figure 2 have now been updated with more defined arrows that point out GLP1R positive cells. The saline vehicle control aortas did not express GLP1R and worked as negative control tissue. Omitting the primary antibody further served as a negative control for the antibody staining. This was now added to the method section, page 3, row 109-111.

4. The numbers in table 1 do not add up. Example: of 22 mice 2 died of AAA rupture leaving 20 lefts. Though only 18 is reported to survive whereof 5 out of 13 developed and AAA. Could you please clarify this.

Reply: We agree that this is not clear. We have double checked our results and updated the table accordingly. The text refers to the first seven days of ruptures while the table 1 refers to total ruptures until day 28.

5. The nicely performed Mass spectrometry imaging is performed on the samples collected on day 7 or on day 28 from the prophylactic trial?

Reply: This is performed on samples collected on day 28. This was now added to the method section page 4, row 133.

6. Please describe how you measured wall thickness in the histological samples. If you have a more representative micrograph of the Ang II + vehicle treated group, I recommend using another Ang II vehicle micrograph as the quality is poor in fig 4 C.

Reply: An average vessel wall area was calculated from the inner luminal diameter and outer adventitia diameter. This was now added to the material and method section, page 3, row 109-111.

7. For statistical analyses of AAA growth and RNA levels. Why do you mainly compare saline control with AngII semaglutide, rather than compare Ang II controls and AngII treated with semaglutide, that to me would be more relevant.

Reply: We agree that the comparisons in the angII groups is more relevant. We performed statistical test between all the groups with multiple test adjustment (post hoc analysis) Most of the alterations were observed between the saline control and angII+semaglutide or saline control and angII+vehicle controls. We did not find many differences between the angII+vehicle vs angII+semaglutide, which is why it looks like we mainly compare saline controls with angII+semaglutide.

8. In the discussion, page 11 line 326 you write “at mRNA levels of col1A1 and Col3A1.. were increased both I the vehicle controls and semaglutide treated group”. Data in table 2 does not support this.

Reply: We thank the reviewer for pointing this out. The * has falled out from the table and has now been added again so the data support this statement.

Minor comments.

For the description of the prophylactic trail where semaglutide was initiated on day 7 continued to day 28 and not for 28 days.

Reply: We are thankful for that the reviewer observed this detail. This has now been changed in the method, page 2, row, 84 and is true for the rescue trial that started with semaglutide on day 7. Both the prophylactic trials (7 and 28 day) started at the same day of disease induction.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I have no further comments.

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Ramilo et al,

Thank you for updating the paper according to my comments. It is a pity that you are not allowed to follow AAA growth by ultrasound because of your animal protocol, then you miss important information.

I find it a bit misleading not clearly stating that no effect of treatment was seen between vehicle and semaglutide treated group and the focus on difference in comparison to saline control condition.

In your table 2 please double check data is correctly included for the SD for MMP2 and MMP9, for me SD =1.00 in both is a very large variation and with the low sample numbers I find it hard that the values are significant different in the Ang II treated groups, however that MMP9 is elevated in this model has been shown many time previously so I would have assumed it would be elevated.

I have no further comments.