Chapter2:Compounds Targeting OSBPL7 Increase ABCA1- Dependent Cholesterol Efflux Preserving Kidney Function in Two Models Of Kidney Disease

For more info. contact tina.xiang@wecistanche.com

Discussion

The present study describes the identification of a class of compounds by phenotypic drug discovery (PDD) that upregulate ABCA1-dependent cholesterol efflux. Target deconvolution revealed OSBPL7 to be the molecular target. In animal models of kidney disease, these compounds reduced renal lipid accumulation, prevented podocyte loss, normalized proteinuria, and reduced renal function decline.

Current therapies for chronic kidney disease are limited to ACEi, ARBs, and, more recently, sodium-glucose cotransporter-type 2 inhibitors(SLGT2i). Despite the effectiveness of these agents in delaying the progression of chronic kidney disease, primarily in patients with proteinuric kidney disease, the unmet medical need in CKD remains immense3,3940.In a randomized trial of 4300 patients, SGLT2i were unexpectedly found to be efficacious regardless of the presence of diabetes4, suggesting that mechanisms of renoprotection other than glucose-lowering are involved. In fact, a recent study demonstrated that SGLT2i can increase plasma HDL and may reduce cardiac fat accumulation42. This recent emergence of SGLT2i underscores the benefit of targeting multiple CKD driving mechanisms to develop more effective therapies.

We, and others, have described glomerular cholesterol accumulation in renal diseases arising from both metabolic and nonmetabolic origins. Glomerular lipid accumulation, in association with decreased ABCAl-mediated cholesterol efflux, occurs in DKD7-11, FSGS12-14, and AS13. These data suggest that restoring ABCAl function may be effective in treating these types of renal disease. We recently demonstrated that renal cholesterol depletion with cyclodextrin, or genetic ABCAl overexpression, protects from renal dysfunction in models of DKD9, FSGS13,14, and ASl3. We also demonstrated that podocyte-specific ABCAl deficiency worsens the renal phenotype in DKD7. Studies with LXR agonists, such as T1317, have also provided strong support for a key role of ABC Al, which has remained an attractive therapeutic target despite the limitation of these agents.

We searched for agents by phenotypic drug discovery and optimization of candidate small molecule drugstore. We identified a series of 5-arylnicotinamides, several representatives of which increased ABCAl protein levels and cholesterol efflux from cholesterol-loaded cells. The absence of efflux activity in fibroblasts from a patient with Tangier disease, and the absence of an increase in ABCAl mRNA, revealed a new mechanism of action.

Target deconvolution efforts identified OSBPL7 as a high-interest target. The molecular probe（³H-Cpd K） efficiently bound OSBPL7 in living cells and could be effectively competed with unlabeled compounds. More importantly, we demonstrated a correlative "activity-activity" relationship for a larger set of compounds between their ability to increase ABCAl cholesterol efflux and their ability to compete for binding of the probe to OSBPL7. Mutational analysis, combined with molecular modeling and docking studies, indicates direct interaction of the compounds with the predicted OSBPL7 sterol binding pocket. Lastly, siRNA studies confirmed the involvement of OSBPL7 in ABC Al function.

Click to know cistanche where to buy and what is cistanche used for  

We also identified CBIR as a potential target. Several reports suggest a role for endocannabinoid signaling in cellular cholesterol regulation. Rimonabant was shown to modestly upregulate scavenger receptor B1 (SR-B1) and ABCGl, but not ABC Al, consistent with our data45. The synthetic cannabinoid WIN55,212-2 was shown to reduce the expression of ABC Al while increasing scavenger receptor CD36 in oxLDL-loaded RAW264.7 macrophages46. Together these data do not support a strong role for CBIR in the regulation of ABCA1.In our studies, rimonabant and other CBIR agonists/antagonists including AM251, WIN55-212-2, and arachidonic choloroethylamine (ACEA), were inactive to induce ABCA1 cholesterol efflux (Figs. 3a, 4a).

Mammalian oxysterol binding proteins(OSBPs)were originally implicated in the biosynthesis of cholesterol and sphingomyelin47-49. Several studies have linked specific OSBPs with the regulation of ABC Al protein stability and ABCAl-mediated cholesterol efflux50,51. No data prior to the present report has implicated OSBPL7 as a regulator of ABCA1. Our data suggest that OSBPL7 may provide a chaperone or substrate transfer function affecting ABC Al. Preliminary co-immunoprecipitation experiments in transfected cells did not identify a direct interaction between the two proteins. However, a meta-analysis of genome-wide association studies (GWAS) identified OSBPL7 among 60 genetic loci influencing plasma cholesterol levels in humans.

Our long-term interest in the role of ABC Al in kidney disease led us to evaluate the 5-arylnicotinimides in vitro and in vivo models of renal disease. We found that OSBPL7 is expressed in kidneys, including in glomeruli and podocytes, and that Cpd A and Cpd G upregulate apo Al-mediated cholesterol efflux from cholesterol-loaded podocytes. Cpd A and Cpd G promote an increase of ABCAl at the plasma membrane without affecting mRNA expression. This mechanism contrasts with LXR agonists, which upregulate ABCAl mRNA leading to an increase in ABC Al in both plasma and intracellular membranes. We evaluated the therapeutic efficacy of Cpd A and Cpd G in vivo in a mouse model of FSGS (ADR-induced nephropathy). Both agents were found to reduce renal cholesterol content, correlating with reduced proteinuria, weight loss, and renal pathological changes. Cpd G was highly effective to prevent ADR-induced podocyte loss and glomerular structural changes. The significant reduction of ABCAl in kidneys, accompanied by an increase in inflammatory proteins, was prevented by Cpd G, supporting a key role for ABCA1. Cpd G effectively reduced albuminuria, BUN, and serum creatinine levels in AS mice. More importantly, Cpd G prolonged lifespan in older Col4a3 animals with advanced renal failure.

These data support the therapeutic potential of compounds targeting ABCAl and strongly suggest that this new mode of action will offer additional, or complementary efficacy to drugs targeting RAAS, SGLT2, NRF2, and cholesterol synthesis or absorption. Notably, in contrast to the effect of Cpd G, RAAS blockade does not prolong lifespan in older Col4a3 mice, despite being the most effective therapy for early intervention in this model53,54. The antiproteinuric activity of Cpd G was superior to the effects of ramipril we previously reported in a syngeneic model of AS38. Cpd G may therefore better target the population of patients with AS and proteinuria when compared to bardoxolone, an NRF-2 agonist that was shown to increase GFR but had no effect on proteinuria, the most powerful predictor of 10-year ESRD risk in the general population55,56. Statins were found to be protective in experimental AS57, and are recommended in all CKD patients to reduce cardiovascular risk, but have proven ineffective to slow the progression of kidney disease 8. Lastly, we recently reported a potential benefit of ezetimibe in AS by reduction of fatty acid and triglyceride absorption. However, ezetimibe did not affect kidney cholesterol content, and its efficacy to improve GFR and reduce proteinuria was inferior to the effects we report here for Cpd G

In summary, the present report describes the identification by phenotypic drug discovery of first-in-class drug-like small molecules that induce ABCAl-mediated cholesterol efflux. Chemical biology identified OSBPL7 as the molecular target and demonstrated that the molecules interact intimately within its predicted oxysterol binding domain. To our knowledge, this is one of only a few examples of drug-like compounds initially discovered by phenotypic approaches, whose target was subsequently successfully identified44. Compounds G and A were highly efficacious to improve renal function in mouse models reflecting two diverse etiologies of progressive renal disease. They appear to be safe and well-tolerated, as reflected by their ability to reduce AST levels in ADR-challenged mice and, more compellingly, to prolong the lifespan of Col4A3 KO mice. These compounds may offer a potential new therapy for renal disease by targeting cellular cholesterol metabolism through a new mode of action. These agents, along with current drugs targeting blood pressure or glucose control, may address the significant unmet need in CKD.

Methods

Compounds. The compounds used in this study are summarized in Supplementary Table 5 and were synthesized at F.Hoffmann-La Roche Ltd, Basel, Switzerland unless otherwise indicated. For in vitro experiments, lyophilized compounds were reconstituted in DMSO (Sigma) and aliquots were stored at -20 ℃. For in vivo experiments, lyophilized compounds were resuspended in vehicles containing 1.25%hydroxypropyl methylcellulose,0.10% docusate sodium salt,0.18% propylparaben sodium, and 0.02% citric acid monohydrate, at pH 6. A fine particle suspension was generated by three brief pulse sonications performed on ice.

Synthetic chemistry procedures and analytical data. Please refer to the Supplementary Data 1 file.

Chemical biology/candidate target screening Briefly, a plasmid expressing a candidate target of interest fused with a FLAG tag was transfected into HEK293 cells. After 24h, tritiated compound K at a concentration capable of inducing ABCA1-mediated cholesterol efflux（1.0 μM） was added for 3h， then the live cells were exposed to UV-light (K=336 nm) to query potential compound/target cross-linking in situ. Cell lysates were prepared, the expressed protein was then selectively immunoprecipitated using an anti-FLAG antibody coupled to sepharose beads, bound proteins were separated by SDS-PAGE, and two identical blots were prepared and probed by western blot and by autoradiography to detect protein expression and presence or absence of compound crosslinking, respectively. The flow chart of this screen is illustrated in Fig. 2. Figure 2d illustrates the ability of Cpd A and Cpd G to displace tritiated Cpd K from overexpressed and immuno-precipitated OSBPL7. Figure 3b illustrates the use of this assay to detect the binding of tritiated Cpd K to overexpressed and immunoprecipitated variants of OSBPL7 containing point mutations in the predicted oxysterol binding pocket of OSBPL7. The advantages of this method are i. the use of plasmid transfection leads to high expression of each candidate which reduces the likelihood to miss an authentic target for those which may be expressed at low endogenous levels, iii. the interaction of the chemical probe with its target occurs in a relevant biological context (live cells) at a pharmacologically active concentration, and iii. immunoprecipitation allows the expressed protein of interest to be purified from background proteins, substantially improving the signal/noise ratio.

Cell culture. Cell lines were purchased from ATCC and subcultured for no more than passage 15, except for human podocytes, which were originally from Dr. Jochen Reiser. Conditionally immortalized human podocytes were cultured under permissive conditions at 33℃, in RPMI medium(Corning) containing 10% FBS (GIBCO),1% penicillin/streptomycin,0.01 mg/ml recombinant human insulin, 0.0055 mg/ml human transferrin(substantially iron-free), and 0.005 ug/ml sodium selenite (ITS, Corning). To induce differentiation, podocytes were cultured at 37 ℃for 14 days in an RPMI medium containing 10% FBS (GIBCO) and 1% penicillin/streptomycin. Differentiated podocytes were starved in RPMI-0.2% FBS for 18h and then treated with 1 uM， 5 uM， and 10 μM of compounds freshly prepared in DMSO for 18 h at 37℃ as indicated.

Cholesterol efflux in macrophages and fibroblasts. The ability of compounds to induce ABCAl-mediated cholesterol efflux was determined in replicate cultures of THP1 cells or fibroblasts in 96-well microplates. Cells were plated at a density of 150,000 cells/well. THP1 monocytes were differentiated to macrophages by the addition of PMA (100 nM) for 72 h in an RPMI medium containing 10% fetal bovine serum and 3 pl of 2-mercaptoethanol. Cells were washed once with RPMI-1640 and then loaded with 50 ug/ml acetylated LDL， and 10μCi/ml 【' H】-cholesterol in RPMI-1640 medium containing 2% FCS, for 48h at 37℃. After loading, cells were washed once with RPMI-1640 and incubated with test compounds for an additional 24h in RPMI-1640 medium containing l mg/ml fatty acid free-bovine serum albumin (BSA). Cells were then washed once, and cholesterol efflux was induced by the addition of 10 ug/ml of human apoAI or 30 ug/ml of HDL2, HDL3, or LDL in RPMI-1640 containing I mg/ml BSA for an additional 6h. Radioactivity in culture supernatants was determined by scintillation counting. Cholesterol efflux was expressed as a relative value vs. controls incubated in the absence of cholesterol acceptors. Dose-response curves were calculated using XLfit3 (ID Business Solutions Ltd. UK).

Cholesterol efflux in podocytes. Human podocytes differentiated for 13 days were labeled for 24h at 37℃ in RPMI medium containing 2% FBS and [PH]-cholesterol (1 uCi/ml, American Radiolabeled Chemicals). Cells were then washed 3 times with PBS and incubated with RPMI supplemented with 0.2% fat-free-BSA (Sigma) with or without compounds for 18-h at 37 ℃. The compounds were used at the following concentrations∶ Cpd C（1 μM）， Cpd A （1 μM， 5 uM）, and Cpd G（l uM， 5 uM，10 μM）. Following the 18-h incubation， human apoAI（Calbiochem）was added to the media, (20 ug/mL final concentration) and the cells were incubated for another 18 h at 37℃. Aliquots of medium collected before(T =0h) and after(T=18h) the addition of apoAI were centrifuged at 12,00×g for 5 min and the radioactivity in the supernatant was counted by liquid scintillation. Cells were then washed with PBS, lysed 0.1% SDS, 0.1M NaOH and the radioactivity in the lysates was quantified by liquid scintillation. apo Al-mediated cholesterol efflux, expressed as %, was calculated as the amount of label released to the medium after adding apo Al (the difference in radioactivity in the medium before and after adding ApoA1) divided by the amount of total label in each well (radioactivity released to the media plus radioactivity in the lysed cells).

siRNA studies. siRNA pools specific for human OSBPL7, human CBIR, and nontarget siRNA negative control are summarized in Supplementary Table 6. THPI monocytes were seeded at a density of 80,000 cells/well in 96-well plates and differentiated with PMA as described above. After 72 h, the medium was replaced by 50 μl of fresh medium per well. Cells were then transfected using Viromer Blue reagent (Origene) following the manufacturer's directions. Briefly,10 ul of siRNA （5 μM） was mixed with 90 μl of Viromer Blue（diluted 1% v/v in supplied buffer）. After 15 min incubation, siRNA-transfection reagent complexes were diluted at 1:6.5 in fresh medium and added to a volume of 50 ul per well. After 24h, cells were then loaded with cholesterol by incubation with acetylated LDL(25 ug/ml) in RPMI medium，1% FBS，100 nM PMA，2 μCi H-cholesterol for 24h. A set of wells containing transfected cells was set aside to verify transfection efficiency by RT-PCR. The cholesterol-loaded macrophages were then incubated in the presence of Cpd G 5 uM, or vehicle (0.1% DMSO) in RPMI medium who Phenol Red, 0.2%BSA) for an additional 24 h. Cholesterol efflux was then measured as described above.

RT-PCR. RNA was isolated using the RNEasy Plus Mini kit (Qiagen). Reverse transcription was performed using qScript cDNA SuperMix(Quantabio) according to the manufacturer's instructions. Quantitative real-time PCR was performed using TaqMan Fast Universal PCR Master Mix and Taqman gene expression assays (Supplementary Table 7).

Western blot. Lysates or cell fractions were incubated in Laemmli buffer at 55℃for 10 min under reducing conditions. A total of 30 ugs of protein from cell lysates or one-third of the volume collected in cell fraction preparations was separated by SDS-PAGE in 4-20% gradient gels (Biorad) and proteins were then transferred to PVDF membranes. After blocking with 5% TBS-milk, membranes were probed with primary antibodies overnight at 4℃. The following primary antibodies were used: mouse anti-ABCA1(Abcam; 1:1000), rabbit anti-GAPDH antibody (Milli-pore;1:10,000),mouse anti-actin(CP01,Millipore 10,000),anti-MEK, anti-ERp72, anti-V5 and anti-Na+/K+ATPase(Cell Signaling; all 1:100), and rabbit anti-OSBPL7(Sigma; 1:1000). After washing, membranes were incubated with anti-rabbit or anti-mouse IgG-HRP antibodies (Promega,1:10,00). Signals were detected after incubation with Westernbright ECL HRP substrate (Advansta)and luminescence signals were captured with Azure C600 Gel Imaging workstation (Azure Biosystems Inc, USA) or X-ray films.

Cell fractionation. Differentiated podocytes were treated with compounds Cpd C （l μM）， Cpd A（5 μM）， or Cpd G（10 μM） for 18 h， washed, and then harvested in ice-cold PBS. Cells were centrifuged at 1000×g for 5 min, the supernatant was carefully removed, and the pellet resuspended in hypotonic buffer(15 mM KCl,1.5 mM MgClz,10mM HEPES, and I mM DTT) supplemented with a protease inhibitor cocktail (Roche). Cells were disrupted with a glass douncer, and sucrose was added to achieve a final concentration of 227mM. Cell lysates were centrifuged at 1000×g for 30 min at 4℃. The supernatant was then transferred to a new tube and centrifuged at 10,000×g for 15 min at 4℃. The pellet, containing the microsomal fraction was collected and the supernatant was transferred to a new tube and centrifuged at 100,000×g for 1 h at 4℃. The pelleted plasma membrane fraction was collected and the supernatant, containing the organelle-free cytosolic fraction, was concentrated in Vivaspin 500 filter columns. Western blot was performed using antibodies for ABCAl and specific markers for each membrane fraction as described above. Nat/K+ ATPase was used to identify the plasma membrane fraction, ERp72 as a marker for the microsomal-enriched fraction, and MEK as the control for the non-membranous cytosolic fraction.

ABC AT co-immunoprecipitation. ABCA1-Flag construct was kindly provided by Dr.Michael Fitzgerald, Harvard University. OSBPL7-V5 construct (pcDNA3.1-V5-ORP7) was purchased from Addgene. Mammalian expression vectors PCMV6-AN-GFP and pCMV6-AC-3DDK, from Origene, were used as negative controls. HEK293 cells were transiently co-transfected with either one of the following plasmid combinations:(1)ABCA1-Flag+OSBPL7-V5;(2)ABCA1-Flag+PCMV6-AN-GFP; or(3) OSBPL7-V5＋pCMV6-AC-3DDK using Fugene 6 transfection reagent (Promega)following the manufacturer's instructions. 48h after transfection, cells were lysed with 1 ml immunoprecipitation buffer (1%Triton X-100,50 mM Tris-Cl pH 7.0,150mM NaCI)supplemented with a protease inhibitor cocktail (Roche). The lysates were cleared by centrifugation and equal amounts of total proteins from each supernatant were incubated overnight at 4℃with 30 μl of anti-Flag M2 Affinity gel （Sigma Aldrich）. The beads were pelleted by centrifugation and washed 5 times with 1 ml of immunoprecipitation buffer. Immunoprecipitated proteins were eluted by incubating the beads with 50 ul of 2x sample buffer for 15 min at 56℃. The presence of ABC A1 and OSBPL7-V5 in the lysates and eluates was verified by western blot as described above.

Animal studies. Animals were housed on 12 h/12h light/dark cycles and under controlled temperature(18-23℃) and humidity(40-60%) in a pathogen-free animal facility of the Division of Veterinary Resources at the University of Miami, Miller School of Medicine. Mice were provided a standard 18% protein rodent chow diet and water ad libitum. The study protocol was approved by the Animal Care and Use Committee (IACUC) of the University of Miami, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC). All procedures were performed in compliance with all federal and state ethical regulations, including the Animal Welfare Act (AWA)regulations overseen by the United States Department of Agriculture(USDA) and the Public Health Service Policy on Humane Care and Use of Laboratory Animals (PHS Policy) administered by the National Institutes of Health (NIH), Office of Laboratory Animal Welfare(OLAW). Female Balb/c mice were purchased from the Jackson Laboratories for Adriamycin-induced nephropathy studies. At 8 weeks of age, mice received a single dose of Adriamycin (Sigma-Aldrich, 12mg/Kg) or vehicle(0.9% NaCI) via tail vein injection. Mice were randomly separated into six experimental groups and, starting at 24 h post-ADR injection, treated with either vehicle or different doses of compounds daily via oral gavage for 28 days. The following doses were used: Cpd C (LXR agonist)10mg/Kg; Cpd A 30mg/Kg; Cpd G 100mg/Kg. All animals were sacrificed 28 days post-ADR injection. Col4a3 KO mice (Col4a3tmiDec strain 129X1/SvJ) were purchased from Jackson Labs. Mice were randomly separated into 2 groups. Starting at 4 weeks of age, the mice were dosed by oral gavage once daily with either vehicle or Cpd G(100 mg/Kg) for 4 weeks followed by sacrifice at 8 weeks of age. A second study was performed in Col4a3 KO mice with treatment initiated at 6 weeks of age. This study was performed in order to assess if Cpd G treatment of mice with established renal failure might delay end-stage renal disease and therefore prolong survival. For ramipril treatment, Col4a3 KO mice started receiving ramipril at the age of 4 weeks. Ramipril was added to the drinking water at a concentration that would lead to a dose of 10 mg/Kg/day 8.

Phenotypic analysis of mice. Weekly spot urine samples were collected. Urinary albumin was measured by ELISA (Bethy! laboratories) and creatinine by a colorimetric assay (Stanbio). Albuminuria was expressed as the albumin/creatinine ratio. Mice were anesthetized with Ketamine (90 mg/Kg)/Xylazine(10mg/Kg), blood was collected in heparinized tubes, and animals were then perfused via the left ventricle with 0.9% NaCl solution. Kidney cortexes were carefully excised and separated into samples that were either snap frozen, embedded in OCT, or fixed in 10% formalin for subsequent assessments.

Determination of lipid deposition in kidney cortex sections. OCT embedded kidney cortex samples were cryosectioned at 4 μm thickness and stained with Oil. Red O-isopropanol solution (Electron Microscopy Science, PA) diluted 6:4 in water and then counterstained with Hematoxylin Harris Hg Free (VWR, PA). Sections were then evaluated using a light microscope(Olympus BX 41, Tokyo, Japan).

Determination of cholesterol and triglyceride content. Snap frozen kidney cortex samples were homogenized in 2 mM potassium phosphate buffer, pH 7,(10 pul buffer/mg tissue） using a glass douncer on ice. 100 μl of the homogenate were mixed thoroughly with I ml of hexane: isopropanol (3:2;v.v) per volume of tissue homogenate. The aqueous (pellet) and organic (supernet) phases were separated by centrifugation at 12,000×g for 5 min. Pellets (aqueous phase) were dried and reconstituted in 8 M urea,0.1% SDS,0.1 M NaOH, and total proteins were quantified by the BCA method (Thermofisher). The organic phase was dried by exposure to nitrogen. Lipids were then reconstituted in 100 μl isopropanol∶NP-40 (9:1;v.v). Triglyceride content was assayed using a colorimetric kit according to the manufacturer's instructions (Cayman Chemical USA). Total cholesterol and CE content were assayed using an enzymatic fluorometric method59. For total cholesterol measurement, extracts were diluted 1:100 in assay buffer (100 mM potassium phosphate,50mM NAC,5mM cholic acid,0.1% Triton X-10,pH 7.4)containing cholesterol oxidase(1 U/ml), cholesterol esterase(1 U/ml), horseradish peroxidase（1 U/ml）， and Amplex Red（75 μM）. The reactions were incubated at 37 ℃ for 30 min in a black opaque 96-well plate (Greiner) and fluorescence was measured in a microplate reader(Spectramax i3X, Molecular Devices) at 530 nm excitation and 580 emissions.

For CE determination,150 ul of the assay buffer described above, supplemented with bovine liver catalase(45 U/ml) and cholesterol oxidase(1 U/ml), was added to 25 μl of sample and incubated overnight at 37°C to deplete free cholesterol leaving only cholesterol esters. Then,75 ul of cholesterol ester detection reagent (1 U/ml cholesterol oxidase, 4 U/ml cholesterol esterase, 24 U/ml horseradish peroxidase, 300 μM Amplex Red） was added. The reaction was incubated at 37℃ for 30 min and the fluorescence signal was measured as described above. Internal standards included samples containing 1 mM of cholesterol and 5 uM of cholesterol oleate in order to verify the completeness of the enzymatic decomposition of free cholesterol as well as assay sensitivity and specificity.

Histological evaluation. Formalin-fixed paraffin-embedded kidney cortex sections 4 um thick were stained with Periodic acid-Schiff(PAS). The slides were analyzed by a renal pathologist in a double-blind design. Glomerular sclerosis, (global and segmental), podocyte hypertrophy, and podocyte hyperplasia were expressed as the percentage of glomeruli found with these abnormalities. Tubular microcysts and interstitial inflammation were scored on a 0-4 scale where 0=0%,05=1-10%,1=11-25%,2:26-50%,3:51-75%,and 4=more than 75% damage. The mesangial expansion was scored based on semi-quantitative analysis(scale 0-5) or percent of glomeruli exhibiting mesangial expansion(%).

Immunofluorescence microscopy. For the detection of ABC Al expression, 4 um paraffin kidney cortex sections were deparaffinized and rehydrated. Antigens were retrieved by boiling for 30 min in citrate buffer pH 6.0(Target Retrieval Solution, Citrate pH 6, Sigma, USA). The sections were permeabilized for 30 min with 0.1%Triton X-100 in TBS. Tissue sections were incubated with 10% goat serum (cat #G9023, Millipore Sigma) and 1% BSA (cat# SP-5050-500, Vector, USA) in TBS-T for 1hr at room temperature to block any non-specific binding. Samples were then incubated overnight at 4℃ with Rat anti-ABCAl(Novus,1:200) and Rabbit anti-WI1 (Abcam,1:200) antibodies in a humidified chamber. After washing extensively, samples were then incubated for I h at RT with fluorochrome-conjugated secondary antibodies (Invitrogen,1:500). Antifade mounting media containing DAPI (Vectashield Plus) was added to the sections before being analyzed

Images were acquired using a FluoView FV1000 Confocal Microscope (Olympus) with an x63 oil objective lens in different planes using a Z-series pattern with a step size of 0.18 um. During analysis, individual planes were deconvoluted and stacked to produce a maximum projected image using flu view software version 4.3b(Olympus). A minimum of 5 images were captured from 5 different fields per sample. Quantification of ABCAl glomerular mean fluorescence intensity (MFI)was done using Fiji—ImageJ software.

For the detection of OSBPL7,4 um frozen kidney cortex sections embedded in OCT, were fixed with PFA-sucrose(2%,4%)in PBS for 20min at RT and permeabilized with 0.3% Triton-PBS for 15 min and blocked with blocking buffer (5% FBS in PBS) for 1h. Tissue sections were then incubated with rabbit anti-OSBPL7 (Sigma, 1:50) and Goat anti-Synaptopodin(Santa Cruz,1:800) antibodies in a humidified chamber overnight at 4℃. After washing extensively, samples were then incubated with fluorochrome-conjugated secondary antibodies, and covered with mounting media as described above. Images were captured using a Leica DMI 6000B fluorescent microscope.

Statistical analysis. GraphPad Prism ver 6 or 7 was used to perform all statistical analyses. Groups were compared using a two-tailed t-test or One-way ANOVA followed by Dunnett's or Tukey's test. F, Brown-Forsythe, or Bartlett's test was used to determine significant variances between the groups, if P<0.05, then groups were compared using two-tailed Mann-Whitney, or Kruskal-Wallis followed by Dunn's tests. The correlation between albuminuria and accumulation of cholesterol ester in kidney cortexes was assessed using the Spearman coefficient test. P values<0.05 were considered statistically significant. Only data from independent experiments were analyzed.

All animals were randomly assigned to treatment groups.