Liver-target nanotechnology facilitates berberine to ameliorate cardio-metabolic diseases

Cardiovascular and metabolic disease (CMD) remains a main cause of premature death worldwide. Berberine (BBR), a lipid-lowering botanic compound with diversified potency against metabolic disorders, is a promising candidate for ameliorating CMD. The liver is the target of BBR so that liver-site accumulation could be important for fulfilling its therapeutic effect. In this study a rational designed micelle (CTA-Mic) consisting of α-tocopheryl hydrophobic core and on-site detachable polyethylene glycol-thiol shell is developed for effective liver deposition of BBR. The bio-distribution analysis proves that the accumulation of BBR in liver is increased by 248.8% assisted by micelles. Up-regulation of a range of energy-related genes is detectable in the HepG2 cells and in vivo. In the high fat diet-fed mice, BBR-CTA-Mic intervention remarkably improves metabolic profiles and reduces the formation of aortic arch plaque. Our results provide proof-of-concept for a liver-targeting strategy to ameliorate CMD using natural medicines facilitated by Nano-technology.


5)
and Figure 5, the authors used InsR, LDL-R and p-AMPK as the readout of the effects of BBR-C TA-Mic by using confocal laser scanning microscopy, western, Flow cytometry and RT-PC R. RT-PC R is not appropriate to detect p-AMPK. Instead, the authors should detect total AMPK using western blotting. In addition to InsR expression, insulin signaling such as p-AKT should also be detected by western blotting. As different methods all demonstrate the expression of these proteins, I would suggest some data such as flow cytometry which is not commonly used for these protein detection, can be moved to supplement to condense the figures and improve readability. Figure 5, HFD-fed mice were treated with BBR-S group (BS, 50mg kg-1 of BBR), BBR-C TA-Mic group (BM, 50mg kg-1 of BBR), and other indicated controls. Did the author compare BBR concentration in plasma, liver and adipose tissue among different groups in this HFD model? since they observe the effects of BBR in liver and adipose. 7) For Figure 5- Figure 8, the ideal control should be empty micelle group (EM) to exclude the nonspecific effects of micelle itself. Significance should be calculated between BBR-C TA-Mic and empty micelle group. 8) The phenotypes of BBR-C TA-Mic treatment in mouse models should be reorganized and condense the figures more logically. I would recommend put plasma parameters in Figure 5A, and body weight, tissue weight in Figure 6AB, and liver HE staining in Figure 7AB together as one figure. Molecular data such as Figure 5BC DE as a single figure. Put all the inflammation related data together as one Figure. Figure 8D can be moved right after Figure 8A to describe IF staining of TNFa and IL6 in liver and adipose tissue, same as the arrangement of western blotting and RT-PC R. Figure 6C could be moved to supplement. Again, Figures should appear in the same order to the description in the main text. 9) Figure 7A-B, please also quantify liver TG content and red O oil staining. 10) One of the conclusion of the study is the potential effects of BBR-C TA-Mic on treating C MD. The authors compared oil-red staining images of aortic arches and diverging blood vessels in all groups in Figure S4. Please also the quantify this data to show the significance.

6)
11) The writing of the manuscript should be improved to be more logic, clear and concise. As mentioned above, Figures should appear in the same order to the main text to make it easier to understood. For instance, according to the description in the results, Fig.S1BC should move after 12) Line 339, Figure 3ABC D, each should be described clearly. 13) Line 405-428, Line 441-461, Line 500-510, Line 513-522 should move to discussion, to make the results section condensed and easier to read. 14) Figure labels using A-abcd are too complex. Some of the data can be condensed or moved to supplement.
Reviewer #2: Remarks to the Author: The MS by Guo et al on Liver-target Nanotechnology Facilitates Berberine to Ameliorate C ardiometabolic Diseases is an extensive study on the positive effect of berberine micelles on various aspects of metabolic disease in the liver. My prime concern is the narrative of the study that the micelles passing intact through the intestine reach the plasma and ultimately release their cargo in the liver. Although tempting to speculate that this would be the case, it is notoriously difficult to stably encapsulate drugs within micelles in biological fluids and it is more likley that BBR separates form the micelles on its journay. The micelles might then still have a solubilizing/absorption enhancement effect which could still be valuable but the "smart liver delivery" as suggested by the authors may be less glamorous.
The studies performed do not confirm the narrative.
The release assays are performed in dialysis bags with 14 kD cutoff sizes, this effectively limits interaction with the majority of proteins, especially relevant in the serum (not plasma) incubations. As a result there is no acceptor for BBR inside the dialysis bag effectively limiting transfer. These studies, therefore only underline that the micelles are stable. Indeed when the dialysis-bag permeable GSH in liver homogenate is added and the micelles fall apart release is observed but this is not reflecting smart liver-specific behavior but rather the experimnetal setup.
Similarly, the C aC o permeation of BBR which is in later experiments used as indication that the micelles pass intact over the monolayer, only measures BBR, so this will not tell in what form the BBR passed over the cell layer. Also the results of Rho123 permeation need to be interpreted with caution as interaction of Rho123 with lipidic molecules has been noted.
The subsequent step in the transport to the liver is difficult to imagine. Why would the micelle be able to avoid opsonization and interaction with any other cells type, but only chooses to engage with the hepatocyte? This is difficult to understand. These issues could be tackled with following the carrier as well as following the drug to see if they follow the same fate. Also competition experiments with other cells than the envisioned target cell could be informative to determine how specific the interaction is. Also a mechanism for the micelle behavior is needed. Which pathways (receptors and endocytic pathways) are responsible for micelle uptake by hepatocytes. Are MPS-cells truly unable to interanlize these constructs? In vivo, these studies should be complemented by a liver cell distribution profile to detemrine the specificity of the micelle for hepatocytes over Kupffer cells endothelial cells and stellate cells. Also, for example a study in apolipoprotein E knock-out mice could shed light on the distribution profile.
The biological effects are interesting and in line with successful delivery of large doses of BBR. Of course, to judge the benefit of micellar targeting it would be informative if a dose response (at least an increase of free BBR) could be added to see to what extent the micellar formulation truly changes the characteristics of the drug's performance beyond the 2.5-fold increase in liver uptake.
The language needs imporvement as some synonyms are chosen that are not often used in the field which sometimes results in awkward sentences.
Reviewer #3: Remarks to the Author: The manuscript titled "Liver-target Nanotechnology Facilitates Berberine to Ameliorate C ardiometabolic Diseases" by Guo et al. reports a method for the delivery of BBR to liver by wraping BBR in BBR-loaded cross-link D-α-tocopheryl polyethylene glycol-thiol succinamide micelle. They claim this method effectively accumulate BBR in liver. By inducing the expression of LDL-R, p-AMPK and Ins-R, BBR-C TA-Mic adminstration ameliorates the metabolic disorders and attenuates aortic arch plaque formation. The manuscript is clearly written and the results are well presented.
There are a few specific issues the authors should address by making modifications to the manuscript or by clarifying in their response, after which I would consider this work suitable for publication in Nature C ommunications.
Major comments: Fig. 4C : a. It is shown that the InsR expression levels in the NC and BS groups are much lower than those in Fig. 5D . Please address why the InsR level is low here. b. Please test the level of phosphor-InsR and evaluate the ratio of phospho-InsR to total InsR. c. Please test the expression level of AMPK and evaluate the ratio of phospho-AMPK to total AMPK. Fig. 5D: a. Please test the level of phosphor-InsR and evaluate the ratio of phospho-InsR to total InsR b. Please test the expression of AMPK and evaluate the ratio of phospho-AMPK to total AMPK. Fig. 6B: The adipocyte size in the BS group is larger than that in the NC and MC groups, shown in the panel on the left. However, as indicated in the BAR graph, the adipocyte size in the BS group is smaller than that in the MC group. Please address the why there is the difference.
Minor comments: Page 17, line371: 'RT-PC R ( Figure 4C ) and western blot analysis ( Figure 4D)' should be western blot analysis ( Figure 4C ) and RT-PC R ( Figure 4D).   Answer: Thank you very much for your hard work to review the article and valuable comments. In the past three months, we have done all the experiments you mentioned (see below), aiming for a good quality manuscript. To better address your concerns, we'd like to briefly introduce the design of the delivery system. Berberine (BBR) is a lipid-lowering drug discovered by our group in 2004. 1 In the past decade, BBR has drawn increasing attention for its bioactivity of safely improving energy metabolism in patients with metabolic syndrome. For years, the challenging question came from BBR's unique pharmacokinetic characteristics, as the absolute bioavailability of BBR (oral) is very poor. 2, 3 By principle, an active-site accumulation could be important for therapeutic reagents to execute their pharmacological effects, and liver is the main target of BBR. 1, 4, 5 Therefore, we propose that a delivery system that can mediate a specific BBR liver-deposition might be a new strategy to enhance BBR's efficacy. Conventional micelles (such as TPGS-Mic) could increase the permeability and bioavailability of BBR. 6 However; the use of TPGS-Mics is not successful, as they are unstable in the GI tract and circulation, thus unable to take the drug for active-site accumulation. 7 In this study, CTA-Mic was designed and developed to keep the advantages of the conventional TPGS micelles and increase the stability. As shown in the results, BBR-CTA-Mic could increase the accumulation of BBR in the liver, rather than in other organs ( Fig. 3; the possible mechanisms are discussed in Q & A 4).
We consider the results of interesting for future formula optimization, aiming for a high efficient BBR treatment. Q: 1) In Figure1C-d, although BBR-CTA-Mic showed a robust BBR release in the exposure of 20% hepatic homogenate, however, compared with BBR-S or BBR-TPGS-Mic, BBR-CTA-Mic actually showed relatively low liver release of BBR. Please explain. It seems inconsistent to better liver uptake of BBR compared with BBR-S in Figure 2.
Answer: Yes, the Figure 1C  2) In Figure 1D, the author demonstrated that BBR-CTA-Mic showed better trans-epithelial drug transportation than other groups in vitro. What's the difference of the plasma amounts of BBR-CTA-Mic and BBR-TPGS -Mic after oral administration to demonstrate more BBR entering circulation through intestinal barrier? Answer: As suggested, the plasma BBR level was compared between BBR-CTA-Mic (orally) and BBR-TPGS-Mic (orally). The result showed that the plasma amount of BBR is higher in the BBR-CTA-Mic treated mice than that in the BBR-TPGS-Mic treated ones. The new data has been included in the revised version (Please see page 332-335, Supplementary Fig. 8).
3) In Figure 2 [11][12][13] In the revised version we have discussed the liver-accumulating mechanism of the BBR-CTA-Mic. (Please see Line 319-328) 5) Figure 4 and Figure 5, the authors used InsR, LDL-R and p-AMPK as the readout of the effects of BBR-CTA-Mic by using confocal laser scanning microscopy, western, Flow cytometry and RT-PCR. RT-PCR is not appropriate to detect p-AMPK. Instead, the authors should detect total AMPK using western blotting. In addition to InsR expression, insulin signaling such as p-AKT should also be detected by western blotting. As different methods all demonstrate the expression of these proteins, I would suggest some data such as flow cytometry which is not commonly used for these protein detection, can be moved to supplement to condense the figures and improve readability. Answer: As suggested, we did the experiment. We have done the comparison of BBR concentration in the plasma, liver and adipose tissues among groups in the HFD feeding C57 mice model. The results showed that, BBR level in adipose of the BBR-CTA-Mic treated mice was higher than that of the BBR-S treated ones, but the increase of BBR in adipose by BBR-CTA-Mic is much less than that seen in liver tissues, suggesting its advantage of accumulating BBR in liver. The results are shown below.
We did not add the figure in the revised version, because 1) the results were very similar to that done with normal mice (Fig. 3B), and 2) the manuscript already has 15 supplementary figures. However, the opinion of the reviewer will be fully respected.  Figure 5A, and body weight, tissue weight in Figure 6AB, and liver HE staining in Figure 7AB together as one figure. Molecular data such as Figure 5BCDE as a single figure. Put all the inflammation related data together as one Figure. Figure   8D can be moved right after Figure 8A to describe IF staining of TNFa and IL6 in liver and adipose tissue, same as the arrangement of western blotting and RT-PCR. Figure   6C could Answer: The manuscript is written in the way that the Results part is together with the Discussion, so that we could discuss issues right after the results presentation.

Bio-distribution Evaluation
However, if the reviewer considers that the Results part should be separated from the Discussion in this paper, we will do it.
Based on these, we propose that a delivery system that can mediate a selective BBR liver-deposition might be a new strategy to enhance BBR's efficacy. We agree with the point that conventional micelles (such as TPGS-Mic) could increase the permeability and bioavailability of BBR; but the use of TPGS-Mics is not successful, as they are unstable and collapsed in the GI tract and circulation, thus unable to achieve their target organ in the intact form. 12,13 In this study, CTA-Mic was designed and developed to keep the advantages of the conventional TPGS micelles and increase the stability. In the novel system, ester could help it to evade the elimination by reticular-endothelial system in liver (such as the Kupffer's cells). [17][18][19] We consider the results of interesting for future formula optimization, aiming for a high efficient BBR treatment.

1.
The release assays are performed in dialysis bags with 14 kD cutoff sizes, this effectively limits interaction with the majority of proteins, especially relevant in the serum (not plasma) incubations. As a result there is no acceptor for BBR inside the dialysis bag effectively limiting transfer. These studies, therefore only underline that the micelles are stable. Indeed when the dialysis-bag permeable GSH in liver homogenate is added and the micelles fall apart release is observed but this is not reflecting smart liver-specific behavior but rather the experimental set-up. Answer: We agree with the concern.
The goal of the assay was to test the organ-selective BBR release mediated by vectors. The four organ mimic environments tested in this project were stomach, intestine, blood and liver, representing the physiological steps after BBR-CTA-Mic oral administration.
The dialysis bag release test was designed to have BBR-CTA-Mic inside the bag and organ mimic environment outside the bag, in attempt to learn the release of BBR influenced by the organ environment. We thought that after exposure to the organ mimic environment (outside bag), BBR-CTA-Mic interacts with the organ components that have proper size to penetrate into the bag, and might cause structure change of the micelle, resulting in BBR release to the outside bag environment. We first started with 14KD bag, as it is often used in releasing test for formula investigation, and found that of the 4 organ mimic environments, liver environment is the one that caused a micelle structure change and the highest release of BBR (to outside bag).
We did not go for large molecule cut-off bags, as the 14KD bag already showed selective release of BBR in liver. In the revised version we have described this design in the Method section, to clarify the principle of the experiment. (Please see Line 621-630) As the comment is so much valuable in the view of environment-micelle interaction (different molecule sizes) we will use dialysis bags with different cut-off in future investigation.
The liver-selective release of BBR-CTA-Mic is demonstrated by the in vivo organ distribution assay (Fig. 3), which showed that BBR-CTA-Mic increased BBR liver accumulation. The possible mechanisms are discussed in Q & A 3).
We have corrected "simulated plasma" to "simulated serum" in the revised with lipidic molecules has been noted. Answer: As instructed, ultra-centrifugal filters (a new experiment) have been used to evaluate the integrity of micelles after the transportation across the monolayer, with a method described by Johnsen et al. 20 The result showed that, more than 50% of We agree with the suggestion on the Rho123's results. Rho123 is a P-gp substrate and commonly used to investigate the activity of P-gp on drug transportation. [21][22][23] We have been very careful in the interpretation of the Rho123-related results.

3.
The subsequent step in the transport to the liver is difficult to imagine. Why would the micelle be able to avoid opsonization and interaction with any other cells type, but only chooses to engage with the hepatocyte? This is difficult to understand.
These issues could be tackled with following the carrier as well as following the drug to see if they follow the same fate. Also competition experiments with other cells than the envisioned target cell could be informative to determine how specific the interaction is. Also a mechanism for the micelle behavior is needed. Which pathways (receptors and endocytic pathways) are responsible for micelle uptake by hepatocytes. Are MPS-cells truly unable to interanlize these constructs? In vivo, these studies should be complemented by a liver cell distribution profile to detemrine the specificity of the micelle for hepatocytes over Kupffer cells endothelial cells and stellate cells. Also, for example a study in apolipoprotein E knock-out micecould shed light on the distribution profile. Answer: We consider the question important to improve the quality of the manuscript, and have added several experiments to address the issues. 3

) The fate of BBR and carrier in vivo.
We agree that chasing the carrier and the drug to see their fate in vivo is a very interesting idea. However, we realize that it is a big work, and therefore have had several meetings to discuss the possibility. The experiment might need to label raw materials with isotopes (or fluorescein) followed by chemical synthesis, thus other collaborator teams and administration permission (e.g. isotope) might be needed.
Also, as the experiments might be quite labor-and time-consuming, we hope to have a chance to do it in the near future, if the reviewer agrees. The language needs improvement as some synonyms are chosen that are not often used in the field which sometimes results in awkward sentences. Answer: We agree. In the revised manuscript, we have tried our best to polish the English/grammar and then asked a native English speaker professional to proof read the paper. We hope that the revised manuscript meet the language criteria.
Reviewer #3 (Remarks to the Author): The manuscript titled "Liver-target Nanotechnology Facilitates Berberine to The manuscript is clearly written and the results are well presented.
There are a few specific issues the authors should address by making modifications to the manuscript or by clarifying in their response, after which I would consider this work suitable for publication in Nature Communications.
Answer: Thank you very much for your hard work to review the article and valuable comments.
Major comments: Q4. Fig. 4C: a. It is shown that the InsR expression levels in the NC and BS groups are much lower than those in Fig. 5D . Please address why the InsR level is low here. b.
Please test the level of phosphor-InsR and evaluate the ratio of phospho-InsR to total InsR. c. Please test the expression level of AMPK and evaluate the ratio of phospho-AMPK to total AMPK. Answer: 1) In the Fig. 4, we tested gene expression in the human HepG2 liver cells, but in the 2) We agree. As suggested, we have tested the level of phosphor-InsR and evaluated the ratio of phospho-InsR to total InsR, as well as tested the expression level of AMPK and evaluate the ratio of phospho-AMPK to total AMPK in the revised manuscript.
(Please see Fig. 4  Answer: We have changed LDL-R and Ins-R in Fig. 4 to LDLR and InsR. We have double checked all the abbreviations in the manuscript to make sure the consistence.
Please improve the English by correcting general grammatical errors.
Answer: We have tried our best to polish the English/grammar and then asked a native English speaker professional to proof read the paper. We believe that the general grammatical error has been corrected and the writing has been improved following the reviewer's instruction.

Reviewers' C omments:
Reviewer #1: None Reviewer #3: Remarks to the Author: I have looked at the response letter and the revised manuscript by Guo and colleagues, and found most of the points raised in the previous round of review have been satisfactorily addressed.
There is a minor issue the authors should address by clarifying in their response, after which I would consider this work suitable for publication in Nature C ommunications.
The minor issue is about Figure 4C . In the BBR-C TA-Mic treated group, we can see strong activation of InsR. However, the insulin signaling was not stimulated because phosphorylation of AKT was not induced.
Reviewer #4: Remarks to the Author: In the revised manuscript, the authors have addressed a number of concerns raised by Reviewer #2 during the initial review. The newly added experiments, such as using ultra-centrifugal filter to evaluate the integrity of the micelles after transportation across the C aco cell layer and dosedependent effect of BBR-C TA-Mic on gene expression in the liver, are useful additions and address some of the technical concerns.
On the other hand, one major issue does remain. The authors claimed that BBR-C TA-Mic had preferential engagement in hepatocytes other than Kupffer cells, endothelial cells and stellate cells. This is in contrary to common observations. Liver sinusoidal endothelial cells (LSEC s) and

Reviewers' comments:
Reviewer #3 (Remarks to the Author): I have looked at the response letter and the revised manuscript by Guo and colleagues, and found most of the points raised in the previous round of review have been satisfactorily addressed.
There is a minor issue the authors should address by clarifying in their response, after which I would consider this work suitable for publication in Nature Communications.
The minor issue is about Figure 4C. In the BBR-CTA-Mic treated group, we can see strong activation of InsR. However, the insulin signaling was not stimulated because phosphorylation of AKT was not induced.
Answer: Thank you very much for your valuable comments.
We agree. Our result showed that BBR-CTA-Mic treatment activated InsR expression.
However, the BBR formulations didn't increase the expression of AKT and p-AKT in the HepG2 cells cultured in insulin-free conventional medium. This phenomenon is consistent with previous findings that the AKT was activated by BBR only when insulin was present 1  Answer: Thank you for your valuable suggestions. In the past three months, we have tried our best to detect BBR content in different cell types using the method you mentioned.
The result showed that the BBR content in the Kupffer cells of the mice treated with BBR-CTA-Mic was less than that from the mice treated with BBR solution ( Supplementary Fig. 8A), indicating a reduced uptake / elimination of BBR by the kupffer cells in liver. Furthermore, LSECs showed almost no uptake of BBR, in both BBR solution and BBR-CTA-Mic-treated mice ( Supplementary Fig. 8B). In the liver of the BBR-CTA-Mic treated mice, the average BBR content in hepatocyte population was similar to that of Kupffer cells ( Supplementary Fig. 8A-c). As the number of hepatocytes in liver is many times more than that of the other types of cells, the sum of BBR in total hepatocytes (mediated through the CTA-Mic entrapment) should be much more than that in other types of cells.
Regarding the possible mechanism, these results might be interpreted by: 1) The chemical and physical feature of the BBR-CTA-Mic (such as PEG chains on the surface of the carrier, as well as the diameter of the particles is within 20-100 nm arrange) could help it to evade the elimination by reticular-endothelial system in liver (such as the Kupffer's cells) [1][2][3][4][5][6][7][8] ; 2) The increased penetration and accumulation of BBR in hepatocytes was facilitated by CTA-Mics.
We did order NTCP antibody (PA5-80001, Thermofisher) for a better labeling of the hepatocytes in the liver cell suspension. However, we were informed recently that the antibody couldn't be available until March 2019. We consider that the new results in this revised version have addressed the concern. However, if the reviewer insists, we could conduct the experiment when we get the antibody.
We agree that, the expression of "BBR-CTA-Mic had preferential engagement in hepatocytes other than Kupffer cells, endothelial cells and stellate cells" might cause confusion, thus we have modified the expression in the new version. (Please see line 288-298 in revised text, Supplementary information and Supplementary Fig. 8).