Intranasal delivery of pro-resolving lipid mediators rescues memory and gamma oscillation impairment in AppNL-G-F/NL-G-F mice

Sustained microglial activation and increased pro-inflammatory signalling cause chronic inflammation and neuronal damage in Alzheimer’s disease (AD). Resolution of inflammation follows neutralization of pathogens and is a response to limit damage and promote healing, mediated by pro-resolving lipid mediators (LMs). Since resolution is impaired in AD brains, we decided to test if intranasal administration of pro-resolving LMs in the AppNL-G-F/NL-G-F mouse model for AD could resolve inflammation and ameliorate pathology in the brain. A mixture of the pro-resolving LMs resolvin (Rv) E1, RvD1, RvD2, maresin 1 (MaR1) and neuroprotectin D1 (NPD1) was administered to stimulate their respective receptors. We examined amyloid load, cognition, neuronal network oscillations, glial activation and inflammatory factors. The treatment ameliorated memory deficits accompanied by a restoration of gamma oscillation deficits, together with a dramatic decrease in microglial activation. These findings open potential avenues for therapeutic exploration of pro-resolving LMs in AD, using a non-invasive route.

Review report COMMSBIO-21-2870-T In the current study, the authors investigate whether intranasal delivery of pro-resolving lipid mediators could resolve inflammation and ameliorate pathology in an AD mouse model. As endogenous activators of resolution, these LMs gain a lot of attention recently, and have been shown to dampen inflammation in a variety of inflammation-related in vivo models, including AD (Maresin 1, PMID: 31680874). To my surprise, this paper is not mentioned at all, whereas a similar treatment (although in combination with other LMs) was applied in the current study. This should be mentioned as it diminishes the novelty of the study. However, the authors compensate for this by (as said) testing a mixture of LMs (RvE1, RvD1, RvD2, Mar1 The and NPD1) in an AD model, but also by providing more in-depth analysis and exploring a new delivery route, which paves the way to design novel treatment strategies using such a non-invasive route. Overall, this is a well-executed and nicely written study but I do have some concerns that need clarification.
Major concerns 1.The authors should mention/discuss the previous study in which maresin 1 was applied in an AD model in all relevant parts of their manuscript. 2.Animal model: the pathology in this mouse model starts at 2 months and peaks at 7 months according to the literature. Can the authors explain why they start their treatment roughly after 6 months, and continue that treatment for roughly two months? This also holds true for the behavioral and other tests, were they performed at the right time point considering the above? 3.Can the authors provide more details on the used LM's, where methyl ester forms used that are generally more stable in vivo compared to 'normal' LMs? If not, why not? 4.Can the authors explain why the currently used treatment protocol was applied (so 3 times a week). To support that protocol, please provide information on the kinetics of these SPMs in vivo. 5.From a translational point of view, it would have been nice if a control group was included that also received the LMs to see what effect they have on all measured parameters. Can the authors comment on this and explain why this group has been left out? 6.In general, a clear mechanism of action is missing, and I also miss a speculation on this in the discussion section. Many parameters are studied (which is great), but before thinking about a mechanism of action, the first thing we need to know is whether the LMs have a direct or indirect effect. To study that, it would be great if the authors can show that the LMs actually reach the CNS (for example by showing the data as introduced (but not shown?) in the method section: Two AppNL-G-F mice were treated with a mixture of deuterium-labelled LMs. Did the authors find them back in the CNS? In SF3 the authors show endogenous levels of LMs in the CNS, but these are not affected by the treatment, and key cytokines (like IL-10 or TNFa), which are well-known responsive cytokines upon LM treatment, are not different between the treated and non-treated group. These findings at this point together suggest that the tested LMs do not reach the CNS, and therefore might display an indirect role in the observed effects. Please comment on this in the relevant parts of this manuscript (especially the discussion) and provide data (if possible) on this with the deuterium-labelled LMs. 7.In line with that, the authors show WB levels of receptors, but to provide a mechanism of action, it would be better to show the actual IHC to provide spatial information (not for all markers, but maybe a selection for the key receptors based on the working hypothesis on the MoA). 8.Why were no changes observed in Abeta? And in line with that, how can the observed in vivo effects be explained? This again is connected to the MoA which needs more attention.
Minor points 1. Rephrase the following sentence in the abstract (' Resolution of inflammation normally follows neutralization of pathogens; and active response to limit damage and promote healing, mediated by pro-resolving lipid mediators (LMs)' ) which in its current form is difficult to follow. I think the authors here would like to make the connection between chronic inflammation and impaired resolution so please adjust. 2. Please mention in the abstract which specialized pro-resolving lipid mediators were tested and why a mixture was chosen. 3. Introduction section: RvD1 also binds to GPR32, please adjust. 4. Results: the authors sometimes use bar graphs (with error bars), sometimes graphs with individual data points, and sometime a combination of these 2. Please use a similar approach throughout the paper (for example like the figures presented in figure 3). 5. Trem2 western blots are difficult to interpret as multiple (faint) bands appear. Which bands are the correct ones? Can the authors provide better images? 6. The text contains a couple of typo's (for example line 216 iquid should be liquid), please adjust and check the text thoroughly.
4. The authors showed that SPMs minimized microglial activation but did not attenuate the levels of pro-inflammatory cytokines or chemokines. However, in the abstract they stated that SPMs decreased microglial activation and proinflammatory cytokines. Please, correct the abstract. Response: The abstract has been corrected.

5.
The authors should speculate (in the discussion) why did SPMs fail to reduce cytokine levels in the brain despite this lipid mediators were able to reduce microgliosis. This is indeed quite surprising, especially, since many reports show that SPMs are able to silence cytokine expression in vitro and in vivo, including in the CNS. Response: This was indeed a surprising finding. The effects on microgliosis were detected as reduction in Iba1, a protein involved in microglia movement, and the observed reduction may therefore in part be an effect on microglial movement. However, this is probably not the entire explanation since the morphology of the microglia would suggest an overall activation. The data on cytokines indeed reveal significantly higher levels in the App NL-G-F mice than in WT animals, whereas the levels upon treatment with SPMs are somewhere in between for several cytokines. We offer as a consideration that the effects of these LMs may elicit functional redundancy to mediate resolution of inflammation and additional neuroprotective/neuroactive signalling, as described for NPD1. Increasing the statistical power by increased number of animals could have given a statistically significant reduction of cytokine levels. The Discussion has been extended in this regard.

The authors should mention/discuss the previous study in which maresin 1 was applied in an AD model in all relevant parts of their manuscript.
Response: This is now cited in all relevant parts of the manuscript.

Animal model: the pathology in this mouse model starts at 2 months and peaks at 7 months according to the literature. Can the authors explain why they start their treatment roughly after 6 months, and continue that treatment for roughly two months? This also holds true for the behavioral and other tests, were they performed at the right time point considering the above?
Response: The treatment was started at 6 months since some several behavioural tests show that impairment starts at this age. There are inconsistencies in the literature as to the appearance of different behavioural changes, and the length of the treatment was chosen to ascertain that there would be detectable impairment compared to WT animals, and that could be altered by the treatment. In addition, the treatment was prolonged since the behavioural testing took 2 weeks, and then there was an extra week for electrophysiology analysis. Furthermore, the translational relevance for the human clinical situation of a treatment starting at an age when behavioural impairment is apparent exceeds the relevance of treatments starting before these occur.
3. Can the authors provide more details on the used LM"s, where methyl ester forms used that are generally more stable in vivo compared to "normal" LMs? If not, why not? Response: The endogenous forms of LMs were used, not the methyl ester forms, and this has now been clarified in the M & M, with addition of catalogue numbers from Cayman Chemicals.

Can the authors explain why the currently used treatment protocol was applied (so 3 times a week). To support that protocol, please provide information on the kinetics of these SPMs in vivo.
Response: The half-life of SPMs is short, e.g. around 5 hours in plasma for RvD1 as shown e.g. by Yellepeddi et al 2021 (Clin Transl Sci 14:683-691, PMID 33202089), and in order to counteract this we decided to give multiple doses. This has been added to the M & M. It would certainly be better in the future to use more stable forms of the SPMs to reduce the number of administrations.

From a translational point of view, it would have been nice if a control group was included that also received the LMs to see what effect they have on all measured parameters. Can the authors comment on this and explain why this group has been left out?
Response: This is a very important point and the rational was simply to optimize the contrast between treated and non-treated animals with pathology, while at the same time limit the number of animals used. It is indeed very valuable to analyse the effects on WT animals, or e.g. the so called APP-WT, that have humanized APP, but not the 3 FAD mutations introduced in the mouse APP in the App NL-G-F mice. The inclusion of WT mice with the same treatment as the AD model will be important for future planned studies.
6. In general, a clear mechanism of action is missing, and I also miss a speculation on this in the discussion section. Many parameters are studied (which is great), but before thinking about a mechanism of action, the first thing we need to know is whether the LMs have a direct or indirect effect. To study that, it would be great if the authors can show that the LMs actually reach the CNS (for example by showing the data as introduced (but not shown?) in the method section: Two AppNL-G-F mice were treated with a mixture of deuterium-labelled LMs. Did the authors find them back in the CNS? In SF3 the authors show endogenous levels of LMs in the CNS, but these are not affected by the treatment, and key cytokines (like IL-10 or TNFa), which are well-known responsive cytokines upon LM treatment, are not different between the treated and non-treated group. These findings at this point together suggest that the tested LMs do not reach the CNS, and therefore might display an indirect role in the observed effects. Please comment on this in the relevant parts of this manuscript (especially the discussion) and provide data (if possible) on this with the deuterium-labelled LMs. Response: The rational for using deuterium-labelled LMs was to define that intranasally delivered LMs under our conditions do reach the brain. As shown in a new figure ( Supplementary Fig. S4) we could detect small amounts of the deuterium-labelled LMs by LC-MS/MS. The interpretation is therefore that the LMs did reach the brain. A possible explanation for the lack of effect on key cytokines is that the effects seen on behaviour and electrophysiology are not due to their resolving effects on inflammation in the brain, but a direct effect on neuronal functions. Possible mechanisms of the effects observed have now been discussed in the manuscript.
7. In line with that, the authors show WB levels of receptors, but to provide a mechanism of action, it would be better to show the actual IHC to provide spatial information (not for all markers, but maybe a selection for the key receptors based on the working hypothesis on the MoA). Response: This is a valid point that WB data do not give spatial information that can advise on possible mechanism of action. We have performed IHC with antibodies to ChemR23 and BLT1 and in the materials remaining it was not possible to make a definite statement regarding where the changes observed by WB occurred. However, as in the human brain, the mouse ChemR23 staining is found in both neurons and microglia, as indicated in the micrographs shown below (Fig. 1). We have added text in the Discussion with possible implications for mechanism of action. Response: There can be several explanations to the lack of effect on Abeta. One possible explanation is that the effects of the LMs are not mediated via the decrease in Abeta, but a direct effect on microglia, and on neurons mediating the improvement in memory functions. Further discussion on the mechanisms of the in vivo effects is now included in the Discussion.