HDLs extract lipophilic drugs from cells

  1. Adi Zheng
  2. Gilles Dubuis
  3. Maria Georgieva
  4. Carla Susana Mendes Ferreira
  5. Marc Serulla
  6. Maria del Carmen Conde Rubio
  7. Evgeniya Trofimenko
  8. Thomas Mercier
  9. Laurent Decosterd
  10. Christian Widmann

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

High-density lipoproteins (HDLs) prevent cell death induced by a variety of cytotoxic drugs. The underlying mechanisms are however still poorly understood. Here, we present evidence that HDLs efficiently protect cells against thapsigargin (TG), a sarco/endoplasmic reticulum (ER) Ca2+-ATPase (SERCA) inhibitor, by extracting the drug from cells. Drug efflux could also be triggered to some extent by low-density lipoproteins and serum. HDLs did not reverse the non-lethal mild ER stress response induced by low TG concentrations or by SERCA knockdown, but HDLs inhibited the toxic SERCA-independent effects mediated by high TG concentrations. HDLs could extract other lipophilic compounds, but not hydrophilic substances. This work shows that HDLs utilize their capacity of loading themselves with lipophilic compounds, akin to their ability to extract cellular cholesterol, to reduce the cell content of hydrophobic drugs. This can be beneficial if lipophilic xenobiotics are toxic but may be detrimental to the therapeutic benefit of lipophilic drugs such as glibenclamide.

Article activity feed

  1. Version published to 10.1242/jcs.258644
  2. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    We thank both reviewers for their insightful comments and suggestions. We propose to address these as described below.

    Reviewer 1

    **Major points:**

    Point 1

    A logical question comes up and I do not think the authors addressed, in a human body what happens to the extracted drugs after loading on HDLs? This requires some mentioning in the discussion.

    1. This is indeed a good question. We have now added in the discussion what may happen to the HDL-extracted drugs in a whole organism. It reads as follows: The likely fate of HDL-extracted drugs in humans is that they are carried to the liver by HDLs. Scavenger receptors such as SR-BI expressed by hepatocytes can then bind HDLs carrying the extracted drugs allowing the drugs to be taken up by the cells. In hepatocytes, the drugs may be inactivated and excreted in the bile (https://doi.org/10.1016/j.cld.2016.08.001, https://doi.org/10.1161/CIRCRESAHA.119.312617). Point 2

    Is the effect specific to the fully mature HDL molecule or do apo-lipoproteins that compose HDLs have similar effects?

    1. This is an interesting question. Apo-AI is the characteristic and most abundant apolipoprotein found in HDLs. It is however not trivial to compare the activities of ApoAI and HDLs because of the difficulty of producing large amounts of ApoAI. In the present paper, the lowest concentration of HDLs that induces drug efflux is 0.125 mM. As there are about 3 molecules of Apo-AI per HDL molecule, we should use 0.375 (3 x 0.125) mM Apo-AI to see if the Apo-AI content of these HDLs can mediate or mimic the drug efflux capacity of the lipoproteins. About 100 mg of recombinant Apo-AI would be required to make 10 ml of a ~0.3 mM Apo-AI cell culture solution. This is an enormous task requiring substantial time and money investment. We are therefore not in a position to perform this experiment that would be of interest but which is not central for supporting the main message of our manuscript. Point 3

    What are non-SERCA-mediated effects of TG?

    1. The SERCA-independent toxic effects of TG have been shown to be a consequence of mitochondrial dysfunction resulting from the ability of TG to induce mitochondrial permeability transition (DOI: 10.1046/j.1432-1327.1999.00724.x). This is now mentioned in the discussion. Point 4

    Why don't HDLs protect cells from low dose TG despite its removal?

    1. Our data indicate indeed that HDLs do not affect the ability of TG to inhibit SERCA and the low ER stress response that ensues. This can be explained by the fact that very low concentrations of TG inhibit SERCA in an irreversible manner (Ki values of 0.2, 1.3, and 12 nM for SERCA1b, SERCA2b, and SERCA3a, respectively) (DOI:https://doi.org/10.1074/jbc.M510978200). Hence, even though HDLs can remove a substantial amount of TG from cells, the concentration of TG that remains in cells is presumably still sufficient to fully inhibits the SERCA pumps. This explanation is now included in the discussion. Point 5

    Line 144. No information on the siRNA was given (refer to the materials section to guide the reader).

    The siPOOLs we have used correspond, for each targeted gene, to a pool of 30 optimally-designed proprietary siRNAs from Biotech. The company does not disclose the sequences of these siRNAs.

    Minor comments:

    Point 6

    There needs to be an abbreviation section. Make sure that you only abbreviate the terms that are used more than once in the text.

    1. An abbreviation list is now provided. Point 7

    Lines 104, 277, 283 and anywhere else: use TG instead of thapsigargin.

    1. Thank you for noting this. This has now been done. Point 8

    Line 262: you don't have to redefine SERCA.

    1. Done Point 9

    I suggest adding structures of the used drugs.

    1. The structures of the drugs used in this work are now presented in Figure S9. Point 10

    I suggest using a table for the RT-PCR primers. Protein Direction Number Sequence Description NCBI entryh-SERCA2 Fwd #1612 5'ATG GGG CTC CAA CGA GTT AC nucleotides 648-667 of human SERCA2, variant a NM_001681.4

    1. Thank you for this suggestion that we have now followed and that indeed facilitates the reading of the RT-PCR method section. Point 11

    Line 93: DMEM (Gibco; ref 61965-059;) the lot number is missing.

    1. The lot number is now indicated. Point 12

    Line 102: 500'000 (and all other thousand numbers) the apostrophe's place is strange.

    1. We have now removed the apostrophe in numbers. Point 13

    Line 381: cholesterol carriers.

    1. This typo has now been corrected Reviewer #2 (Evidence, reproducibility and clarity (Required)):

    **Major concerns**

    Point 14

    1. Figure 2, The authors should perform western blot to evaluate the protein expression levels (not only mRNA levels by Q-PCR)
    1. We have performed these experiments in the past in MIN6 cells (Pétremand et al. Diabetes 2012 May; 61(5): 1100-1111; Figure 2). This earlier work showed that HDLs reduce the induction of TG-induced ER stress markers at the protein (CHOP and BiP) and functionality (IRE1 activity on XBP1 splicing). We will repeat these experiments in DLD1 cells as per the reviewer’s suggestion. Point 15.
    1. Could the authors evaluate whether HDL treatment reduces the amount of SERCA (mRNA/protein) in their cells? The loss of SERCA could explain the reduced accumulation of the BODIPY-TG in the cell?

    We would argue that it is unlikely that a reduction in SERCA expression from cells has any significant impact on TG cell loading as the cell-associated drug is certainly in vast excess compared to the number of SERCA molecules in cells. We will nevertheless perform the requested experiment using DLD-1 cells and assess whether HDLs modulate their SERCA2 expression.

    Point 16.

    1. To generalize their observation, It would have been interesting to test more lipophilic/hydrophilic drugs to quantitatively validate that HDLs are selective of lipophilic drugs.

    We will test 2 new lipophilic (letermovir and lumefantrine) and 2 new hydrophilic drugs (levetiracetam and cefepime) for their ability to be extracted by HDLs (experiment set-up as in Figure 4).

    Point 17.

    1. The ABC transporter part in this manuscript has to be improved with the down-regulation of extinction of ABCA1 and ABCG1 to determine in a comprehensive manner the effect of these transporters in the pro-survival role of HDL.

    We will invalidate the genes encoding ABCA1, ABCB1, ABCG1, and ABCG2 using the CRISPR/Cas9 technology and test the ability of the invalidated cells to promote efflux of thapsigargin to HDLs (experiment set-up as in Figure 6) and to protect them from the drug (experiment set-up as in Figure 6). The choice of the cell lines to be used for the invalidation depends on what ABC transporters they express. No single cell line expresses all four ABC transporters to high levels. The following cell lines will be used because, according to the literature or to the Human Protein Atlas (https://www.proteinatlas.org/), they display strong expression of the indicated transporters: for ABCA1: HCT116; for ABCB1: HEK293T; for ABCG1 and ABCG2: MCF7. For consistency with the experiments already performed in the manuscript, the invalidation will also be performed in the DLD1 cell line.

    **Minor point:** Point 18.

    1. ABCB1 blot in figure 7B is not convincing and should be improved.
    1. We will redo this WB to improve the quality of the blot.
  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    In this manuscript, Christian Widmann and colleagues describe how HDLs can protect cells by promoting the extraction of lipophilic drugs such as thapsigargin (TG). The authors observe that HDLs do not affect the ability of TG to inhibit SERCA but instead decrease lipophilic drug content inside cells and therefore protect cells against their lethal effects. Using some compounds (probably not enough to conclude), the authors claim that HDLs can promote the exclusion of lipophilic drugs while hydrophilic drugs or compounds like doxorubicin hydrochloride, an anticancer drug, or Rhodamine 123, were not extracted from cells. Finally using small interfering RNA, the authors reveal that ABCB1 mediates some of the drug effluxes to HDLs. This study is sound and well-written. Although of interest from a therapeutic standpoint, this manuscript should address some questions to strengthen these data.

    Major concerns

    1. Figure 2, The authors should perform western blot to evaluate the protein expression levels (not only mRNA levels by Q-PCR)
    2. Could the authors evaluate whether HDL treatment reduces the amount of SERCA (mRNA/protein) in their cells? The loss of SERCA could explain the reduced accumulation of the BODIPY-TG in the cell?
    3. To generalize their observation, It would have been interesting to test more lipophilic/hydrophilic drugs to quantitatively validate that HDLs are selective of lipophilic drugs.
    4. The ABC transporter part in this manuscript has to be improved with the down-regulation of extinction of ABCA1 and ABCG1 to determine in a comprehensive manner the effect of these transporters in the pro-survival role of HDL.

    Minor point:

    1. ABCB1 blot in figure 7B is not convincing and should be improved.

    Significance

    This study can interest a large scientific audience. Some additional experiments have to be performed to render more convincing some part of this study.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    It was my pleasure to evaluate the work submitted to Review Commons. I have reviewed the work and my comments are as follows: This manuscript entitled "HDLs extract lipophilic drugs from cells" by Zheng and colleagues describes a new mechanistic picture of how HDLs protect cells against death. The authors meticulously describe a novel ability of HDLs to extract hydrophobic xenobiotics from cells akin to their cholesterol-extracting function. I would like to thank the authors for a pleasurable read and their well-defined experimental design. This manuscript is of great value and significance to the fields of clinical chemistry and pharmacology. I therefore do think this manuscript merits publication after tending to these major and minor comments.

    Major points:

    • A logical question comes up and I do not think the authors addressed, in a human body what happens to the extracted drugs after loading on HDLs? This requires some mentioning in the discussion.
    • Is the effect specific to the fully mature HDL molecule or do apo-lipoproteins that compose HDLs have similar effects?
    • What are non-SERCA-mediated effects of TG?
    • Why don't HDLs protect cells from low dose TG despite its removal?
    • Line 144. No information on the siRNA was given (refer to the materials section to guide the reader). Minor comments:
    • There needs to be an abbreviation section. Make sure that you only abbreviate the terms that are used more than once in the text.
    • Lines 104, 277, 283 and anywhere else: use TG instead of thapsigargin.
    • Line 262: you don't have to redefine SERCA.
    • I suggest adding structures of the used drugs.
    • I suggest using a table for the RT-PCR primers. Protein Direction Number Sequence Description NCBI entry h-SERCA2 Fwd #1612 5'ATG GGG CTC CAA CGA GTT AC nucleotides 648-667 of human SERCA2, variant a NM_001681.4
    • Line 93: DMEM (Gibco; ref 61965-059;) the lot number is missing.
    • Line 102: 500'000 (and all other thousand numbers) the apostrophe's place is strange.
    • Line 381: cholesterol carriers.

    Significance

    This manuscript is of great value and significance to the fields of clinical chemistry and pharmacology.

    Referees cross-commenting

    I agree with the experiments suggested by reviewer #2

  5. Version published to 10.1101/2020.12.08.415984v1 on bioRxiv