Inhibition of ABCA1 protein degradation promotes HDL cholesterol efflux capacity and RCT and reduces atherosclerosis in mice.

ABCA1 plays a key role in the initial lipidation of apoA-I, which generates circulating HDL cholesterol. Whereas it is known that the transcriptional upregulation of ABCA1 promotes HDL formation and reverse cholesterol transport (RCT), it is not known how the inhibition of ABCA1 protein degradation impacts HDL function. Employing the small molecule triacetyl-3-hydroxyphenyladenosine (IMM-H007), we determined how the attenuation of ABCA1 protein degradation affects HDL cholesterol efflux capacity, RCT, and atherosclerotic lesion formation. Pulse-chase analysis revealed that IMM-H007 inhibits ABCA1 degradation and facilitates its cell-surface localization in macrophages, and additional studies in macrophages showed that IMM-H007 thereby promotes cholesterol efflux. IMM-H007 treatment of Paigen diet-fed mice caused an increase in circulating HDL level, it increased the cholesterol efflux capacity of HDL, and it enhanced in vivo RCT from macrophages to the plasma, liver, and feces. Furthermore, ABCA1 degradation suppression by IMM-H007 reduced atherosclerotic plaque formation in apoE(-/-) mice. Thus, via effects on both ABCA1-expressing cells and circulating HDL function, the inhibition of ABCA1 protein degradation by IMM-H007 promotes HDL cholesterol efflux capacity and RCT and attenuates atherogenesis. IMM-H007 potentially represents a lead compound for the development of agents to augment HDL function.

fasted for 4 h and then bled from the retro-orbital plexus. All procedures were conducted in conformity with the Public Health Service policy, and they were approved by the Animal Care and Use Committee of the Institute of Material Medica, Chinese Academy of Medical Sciences, and Peking Union Medical College (Beijing, China).

ABCA1 degradation analysis
To determine the degradation rate of ABCA1, THP-1 macrophages were incubated with either DMSO (vehicle) or 10 mol/l IMM-H007 for 1 h. Cells were washed with PBS and then maintained in 0.2% BSA medium with 20 g/ml cycloheximide (CHX) for the indicated time periods. To examine calpain-mediated degradation, THP-1 macrophages were incubated with EGTA, 1,2-bis(2-aminophenoxy)ethane-N,N,N ′ ,N ′ -tetraacetate-AM (BAPTA), or calpeptin (calpain inhibitor) plus CHX for 60 min. In other studies, after treatment with EGTA/BAPTA (1 M/100 M) for 1 h, the cells were treated with CHX plus various extracellular Ca 2+ concentrations with/without IMM-H007 for 30 min ( 16 ). ABCA1 protein abundance was analyzed by immunoblot analysis. Intracellular calpain activity was measured using a commercially available kit (Millipore).

ABCA1 labeling and tracing
To investigate ABCA1 abundance on the plasma membrane and the degradation of the cell surface-resident protein, biotinylation was carried out on THP-1 macrophages as previously described ( 13,17,18 ). Briefl y, cells were incubated at 4°C for 60 min with PBS containing EZ-Link sulfo-NHS-SS-biotin (Pierce). After quenching, cells were washed and lysed in RIPA buffer, and biotinylated proteins were isolated by incubating with streptavidin-agarose resin (Pierce) overnight at 4°C. The bound proteins were eluted with loading buffer, and then subjected to immunoblot analysis.
To trace ABCA1 internalization, the biotinylation of the cell surface protein was cleaved with 50 mmol/l reduced glutathione (Sigma) at pH 7.8 twice for 20 min, and the remaining intracellular biotinylated ABCA1 was determined as the internalized portion. To examine ABCA1 recycling, intracellular ABCA1 was prelabeled as above, cells were treated with vehicle or IMM-H007 for the indicated periods, and cell surface biotin was again removed. Resurfaced ABCA1 was estimated by comparing biotinylated ABCA1 with and without the second cleavage ( 19 ).

ABCA1 fl ow cytometry and confocal microscopy
After treatment with vehicle or IMM-H007 (10 mol/l), differentiated THP-1 macrophages were harvested and suspended in PBS. Cells were blocked with normal goat serum on ice for 30 min and stained with an anti-ABCA1 antibody (Abcam) at 1:50 dilution for 1 h on ice. Following washing, cells were stained with green fl uorescent Alexa Fluor 488-conjugated goat anti-mouse secondary antibody (1:100 dilution), cells were rewashed and resuspended in 1% paraformaldehyde, and FACScan analysis was performed with CellQuest software.
Standard protocols for immunofl uorescence microscopy were employed as described previously ( 20 ). THP-1 cells plated on cover glass were treated with or without IMM-H007 overnight, washed with PBS, fi xed with 4% paraformaldehyde at room temperature for 20 min, blocked with normal goat serum, and stained with anti-ABCA1 antibody (1:50) overnight at 4°C. Cells were rewashed twice and incubated with AlexaFluor 488 antimouse IgG secondary antibody for 30 min. Images were captured with an Olympus confocal microscope and analyzed with Olympus confocal software. initial steps of HDL formation and of RCT, with hepatic, intestine, or adipose ABCA1 activity being closely associated with plasma HDL level ( 8 ). Increased ABCA1 expression in macrophages would be anticipated to enhance RCT and reduce atherosclerosis risk. Thus, ABCA1 has been identifi ed as an emerging target for new pharmacological agents designed to leverage the cardiovascular protective potential of HDL. It is well-established that liver X receptor (LXR)-induced transcriptional upregulation of ABCA1 promotes HDL formation and RCT ( 9 ), and a previous study demonstrated that two oxidized products of probucol inhibit ABCA1 degradation, increase HDL level, and reduce atherosclerosis ( 10 ). However, how altering ABCA1 protein degradation impacts HDL function is unknown.
In the present study, we determined whether an increase in endogenous ABCA1 expression due to altered posttranslational regulation leads to enhanced HDL function and RCT, and consequently an alleviation of atherosclerosis formation. We did so using a novel agent, triacetyl-3-hydroxyphenyladenosine (IMM-H007), which we formerly observed impacts lipid levels in hyperlipidemic hamsters ( 11 ). We discovered that IMM-H007 inhibits intracellular ABCA1 degradation, and that this is related to the suppression of calcium-activated calpain activity. The actions of IMM-H007 on ABCA1 result in greater cellsurface ABCA1 content, increased circulating HDL, enhanced HDL cholesterol effl ux capacity, the promotion of ABCA1-mediated RCT, and a reduction in atherosclerotic plaque size in hypercholesterolemic mice. As such, IMM-H007 is a promising lead agent candidate for raising HDL quantity, and likely more importantly, for enhancing HDL function.

Cell culture
THP-1 and J774 cells were obtained from the Cell Culture Center of Peking Union Medical College (China), and maintained as described (12)(13)(14). THP-1 cells were grown in suspension in RPMI 1640 supplemented with 10% FBS, and they were differentiated to macrophages by incubating in the presence of 10 ng/ml of PMA (Sigma) for 48 h. J774 cells were cultured in Dulbecco's modifi ed Eagle's medium plus 10% FBS, 100 U/ml penicillin, and 100 g/ml streptomycin in a 5% CO 2 incubator.

Atherosclerosis analysis
To induce atherosclerosis in apoE Ϫ / Ϫ mice, animals were fed with the Paigen diet (containing 1.25% cholesterol) ad libitum for 10 weeks. Probucol (200 mg/kg) was employed as a positive control intervention. Mouse aorta preparation and atherosclerosis quantifi cation was conducted as previously described ( 29 ). The entire aorta was stained with Oil Red O for en face analysis and photographed with a Nikon D600 digital camera. Aortic roots were frozen in OCT, and 7 m-thick sections were stained with Oil Red O and hematoxylin to evaluate lipid content. Lesion areas were quantifi ed with ImageJ software.

Statistical analyses
Results are presented as mean ± SEM. Differences between means of two groups were evaluated by Student's t -test (two tailed). ANOVA was used to compare fi ndings over time or between three or more groups, with Tukey's or Sidak's post hoc testing. Analyses were performed using GraphPad Prism 5, and statistical signifi cance was assumed for P р 0.05.

IMM-H007 increases ABCA1 protein but not mRNA level in vitro and in vivo
Previous studies have demonstrated that posttranslational processes play a pivotal role in the regulation of ABCA1 expression ( 30 ). Whether IMM-H007, whose structure is shown in Fig. 1A , affects ABCA1 protein abundance was fi rst tested in THP-1 cells. We found that treatment of THP-1 cells with IMM-H007 increased the protein but not mRNA expression of ABCA1 in a time-and concentrationdependent manner ( Fig. 1B-E ). Identical results were obtained in J774 cells (supplementary Fig. 1A-C), and in mouse peritoneal macrophages (supplementary Fig. 1D, E). IMM-H007 did not alter the expression of LXR, which is a key upstream promotor of ABCA1 expression (supplementary Fig. 2A, B). Consistent with the in vitro fi ndings, hepatic ABCA1 expression was increased in 200 mg/kg IMM-H007-treated mice fed either a Paigen diet ( Fig. 1F ) or standard chow (supplementary Fig. 1F), and there were no changes in hepatic ABCA1 mRNA levels in either the treated mice on Paigen diet ( Fig. 1G ) or the treated mice on standard chow (supplementary Fig. 1G). These data indicate

Plasma lipid profi les
Plasma HDL levels were measured by cholesterol enzymatic kit (BioSino Corporation) after precipitation of apoB-containing lipoproteins by adding 100 l plasma to 40 l 20% PEG 8000 (P-2139 in 200 mmol/l glycine, pH 10; Sigma-Aldrich) solution ( 21,22 ). Plasma lipoprotein levels were analyzed by fast protein liquid chromatography (FPLC) using Superose HR6 columns, followed by fl uorescent enzymatic assays for cholesterol.

Fecal neutral sterol determinations
Mice were individually housed and fed a chow diet or Paigen diet with/without IMM-H007 treatment for 14 days. Feces were collected over the last 3 days and were dried, weighed, and ground to a powder. An aliquot of 500 mg of feces was extracted, and the neutral sterol content (cholesterol, stigmasterol, coprostanol, and ␤ -sitosterol) was quantitated by gas chromatography/mass spectrometry as described ( 23 ). 5-Cholestene was employed as an internal control.

In vitro cholesterol effl ux
Cholesterol effl ux experiments were performed as previously described (24)(25)(26). J774 cells or differentiated THP-1 macrophages grown in 24 multi-well plates were labeled with 3 H-cholesterol (2 Ci/ml) and loaded with 25 g/ml acetylated (ac)LDL in DMEM or RPMI 1640 plus 1% FBS for 24 h. Cells were washed and equilibrated overnight in either the presence or absence of IMM-H007 (25, 50, and 100 M) in medium with 0.2% BSA. Cholesterol effl ux was assessed over 4 h in medium plus 0.2% BSA, with various lipid acceptors added, as detailed in the fi gure legends. Radioactivity was measured in the medium and cell lysate, and effl ux was calculated as the percent radioactivity in the medium relative to the total radioactivity in the cells and medium ( 27 ).

In vivo macrophage RCT
Experiments were carried out as described previously ( 24,25,28 ). apoE Ϫ / Ϫ mice were fed a Paigen diet and treated with vehicle or IMM-H007 by oral gavage for 2 weeks. J774 macrophages were loaded with 50 g/ml acLDL and 5 Ci/ml [ 3 H]cholesterol for 24 h, equilibrated in DMEM plus 0.2% BSA overnight, scraped into DMEM/0.2%BSA, centrifuged (1,200 g , 5 min), and resuspended in DMEM. The labeled J774 cells (4.5 × 10 6 cells/ mouse, 3 × 10 6 cpm in 0.25 ml DMEM, n = 5/group) were injected intraperitoneally into individually housed mice. Plasma samples were obtained at 6, 24, and 48 h after injection, and 10 l aliquots were counted in a scintillation counter. Feces were collected continuously from 0 to 48 h and soaked in distilled water overnight at 4°C, homogenized in an equal volume of ethanol, and 20 l aliquots were counted. Mice continued to receive vehicle or IMM-H007 during the 48 h RCT study. At study termination, mice were euthanized, 100 mg liver was removed for lipid extraction, and 10 l aliquots were counted. Gall bladders were also isolated and 20 l of bile was counted. Results were expressed as percentage of counts per minute injected .

Quantitative real-time RT-PCR
To evaluate gene expression in liver or cells, total RNA was extracted using TRIzol reagent and subjected to reverse transcription using a high-capacity cDNA reverse transcription kit (Takara). Quantitative RT-PCR analyses were performed by SYBR green methodologies and relative mRNA levels were calculated using the comparative Ct method ( 20 ). Each sample was tested in triplicate, and transcript abundance was normalized to ␤ -actin mRNA level. Relative transcript expression was determined using a control sample as a calibrator and the ⌬ ⌬ CT method. The following specifi c primers increases after IMM-H007 stimulation ( Fig. 2B ), paralleling the preservation of total cellular ABCA1.
To further investigate the mechanism for the elevation of cell surface ABCA1, the internalization and recycling of ABCA1 were examined. In the presence of the calpain inhibitor, calpeptin, there was decreased ABCA1 internalization (supplementary Fig. 2C), and it was found that IMM-H007 also inhibits ABCA1 internalization ( Fig. 2C ). Furthermore, we observed that internalized ABCA1 gets degraded rapidly in the absence of IMM-H007 and less rapidly in the presence of IMM-H007 (supplementary Fig. 2D), and there is a resulting greater recycling of ABCA1 to the cell surface in the presence of the agent ( Fig. 2D, E ). Calpeptin similarly caused increased abundance of ABCA1 recycled to the plasma membrane (supplementary Fig. 2E). The enhancement of cell surface ABCA1 by IMM-H007 was further verifi ed by fl ow cytometry and confocal imaging ( Fig. 2F, G ). that IMM-H007 promotes ABCA1 protein but not mRNA expression in vitro and in vivo.

IMM-H007 retards ABCA1 protein degradation and facilitates its surface localization
In combination with reports that ABCA1 protein and mRNA expression are discordant ( 31 ), these fi ndings led us to hypothesize that IMM-H007 increases ABCA1 expression by altering the stability of the ABCA1 protein. Using protein synthesis inhibition with CHX, we observed that treatment of THP-1 macrophages with IMM-H007 delays ABCA1 protein degradation ( Fig. 2A ). Next, we determined whether the inhibition of ABCA1 degradation leads to its accumulation on the cell surface, because the activation of ABCA1-mediated cholesterol effl ux and HDL biogenesis take place at the cell surface. Using biotinylation, we observed that cell surface ABCA1 in THP-1 cells  B) and (C), total RNA was isolated, and quantitative RT-PCR was performed. F, G: IMM-H007 increases hepatic ABCA1 protein but not ABCA1 mRNA level. Hepatic ABCA1 protein (F) and mRNA (G) were evaluated in Paigen diet-fed mice administered the indicated doses of IMM-H007 for 2 weeks. Data are mean ± SEM, n = 3 for (B-E) and n = 10 for (F, G). ** P < 0.01, *** P < 0.001 versus control. NS, nonsignifi cant. Fig. 2. IMM-H007 retards ABCA1 protein degradation and facilitates its surface localization. A: IMM-H007 delays ABCA1 protein degradation. THP-1 cells were incubated with CHX (20 g/ml) for the indicated times after 1 h of vehicle or IMM-H007 treatment. Cell lysates were processed for immunoblotting. B: IMM-H007 increases cell surface ABCA1. THP-1 cells were treated with vehicle or IMM-H007 (10 M) for the indicated times, labeled with sulfo-SS-biotin at 4°C, and biotinylated protein was selectively absorbed by streptavidin-agarose and analyzed by immunoblotting. C: IMM-H007 retards surface ABCA1 internalization. After biotinylation, surface ABCA1 was internalized in the presence or absence of IMM-H007 at 37°C for the indicated times. Surface biotinylation was cleaved and the remaining intracellular biotinylated ABCA1 was quantifi ed to evaluate the internalized fraction. D, E: IMM-H007 facilitates the recycling of internalized ABCA1 to the plasma membrane. After internalization at 37°C for 1 h, biotinylated protein in the vehicle-or IMM-H007-treated cells was recycled at 37°C for the indicated times. Remaining intracellular biotinylated ABCA1 (ic) and total biotinylated ABCA1, which includes intracellular biotinylated ABCA1 (ic) plus surface biotinylated ABCA1 (s), after recycling were detected by immunoblotting. The difference between ic and ic+s represents recycled ABCA1. F: THP-1 cell surface ABCA1 was evaluated by fl ow cytometry after treatment with vehicle or 10 M IMM-H007 for 16 h, using anti-ABCA1 primary antibody and green fl uorescent Alexa Fluor 488-conjugated secondary antibody. G: Confocal microscopy was also performed on similarly-treated cells. Images shown display relative fl uorescent intensity with identical exposure time for vehicle-and IMM-H007-treated cells. Data shown are mean ± SEM, for three independent assays in (A), (B), and (E), and in two studies in (C). * P < 0.05 versus vehicle. In (G), scale bar: 10 m.
Thus, IMM-H007 increases cell surface ABCA1 by suppressing ABCA1 degradation and internalization, and by facilitating ABCA1 recycling to the cell surface.

IMM-H007 restrains calpain activity to inhibit ABCA1 protein degradation
Calpain-mediated proteolysis plays a major role in regulating the activity of ABCA1 ( 32 ). We therefore evaluated the impact of IMM-H007 on calpain-mediated ABCA1 degradation in THP-1 cells. As expected, the inhibition of calpainmediated degradation with calpeptin or by chelating of calcium ion increased ABCA1 protein level ( Fig. 3A ), and increasing medium Ca 2+ concentrations accelerated ABCA1 degradation ( Fig. 3B ). IMM-H007 treatment suppressed Ca 2+ -dependent ABCA1 degradation ( Fig. 3B ). Moreover, mirroring the actions of calpeptin, IMM-H007 decreased intracellular calpain activity ( Fig. 3C ), although calpain abundance was invariant (data not shown). The depletion of intracellular calpain expression by siRNA increased ABCA1 protein stability, and IMM-H007 had no further effect on ABCA1 abundance in the setting of calpain silencing. ( Fig. 3D ). These fi ndings indicate that IMM-H007 increases ABCA1 abundance by restraining calpain-mediated ABCA1 protein degradation.

IMM-H007 increases HDL cholesterol and facilitates cholesterol excretion
Next we measured plasma HDL in vehicle-, probucol-, or IMM-H007-treated mice and hamsters. Six weeks of treatment with probucol in Paigen diet-fed apoE Ϫ / Ϫ mice caused a modest 13.2% decrease in circulating HDL, whereas IMM-H007 (100 and 200 mg/kg) resulted in 20.2% and 51.3% increases in HDL ( Fig. 4A ), although plasma apoA-I was not affected (supplementary Fig. 3A). In WT mice fed a Paigen diet, the plasma HDL was elevated 21.3% by IMM-H007 (200 mg/kg) administration for two weeks -dependent ABCA1 protein degradation. Cells were exposed to various Ca 2+ concentrations in the presence or absence of 10 M IMM-H007 at 37°C for 30 min, and analyzed for ABCA1 abundance. C: Intracellular calpain activity is suppressed by IMM-H007. Cells were treated as described in (A), and calpain activity was determined. D: IMM-H007 does not affect ABCA1 stability in calpain-defi cient cells. Cells were incubated in the presence of CHX for 1 h in the absence or presence of IMM-H007 (10 M ) following transfection with control siRNA or siRNA targeting calpain. ABCA1 abundance was then evaluated. In (A-C), data are mean ± SEM of three independent assays. * P < 0.05 and ** P < 0.01 versus control or vehicle.
HDL isolated from IMM-H007-treated chow-fed mice enhanced cholesterol effl ux from J774 and THP-1 macrophages, promoting effl ux up to 1.4 ± 0.75-fold ( P < 0.01) and 1.6 ± 0.53-fold ( P < 0.01), respectively, at the highest dose of agent administered in comparison with HDL from vehicle-treated mice. Regarding ABCA1-mediated cholesterol effl ux ( Fig. 5A, B ), HDL from mice treated with 200 mg/kg IMM-H007 raised effl ux up to 3.18-fold and 6.89fold in J774 and THP-1 cells, respectively. The treatment of J774 or THP-1 cells with IMM-H007 raised cholesterol effl ux to apoA-I ( Fig. 5C, D , respectively). Parallel fi ndings were obtained when primary macrophages were employed (supplementary Fig. 4). In contrast, such treatment of J774 cells did not affect SR-BI-mediated effl ux, which was determined as the component of effl ux to PEG-HDL from vehicle-treated mice that was attenuated by block lipid transport-1 (BLT-1), which is a specifi c inhibitor of SR-BI-mediated lipid transfer ( 33 ) ( Fig. 5E ). The treatment of J774 cells with IMM-H007 also did not affect efflux mediated by aqueous diffusion, which was evaluated by measuring effl ux to methyl-␤ -cyclodextrin ( 12 ) ( Fig. 5F ). Thus, the increased effi ciency of HDL isolated from IMM-H007-treated mice to remove cholesterol from macrophages, and the enhanced effl ux observed following the direct treatment of macrophages with IMM-H0007 are both attributable to more effi cient cholesterol effl ux by ABCA1. ( Fig. 4B ), and the treatment of hamsters with 200 mg/kg IMM-H007 for 14 days raised HDL by 31.3% ( Fig. 4C ) Analysis of lipoproteins in chow-fed WT mice by FPLC revealed that there is an increase in the HDL fraction in IMM-H007 (200 mg/kg)-treated chow-fed mice ( Fig. 4D ). Furthermore, IMM-H007 (200 mg/kg) treatment increased fecal neutral cholesterol excretion in both chow diet-and Paigen diet-fed WT mice ( Fig. 4E, F , respectively). Thus, IMM-H007 raises HDL level and promotes fecal neutral cholesterol excretion. Regarding circulating lipids, both (LDL and triglycerides were decreased in hamsters after 200 mg/kg IMM-H007 treatment for 2 weeks, there was a decrease in circulating LDL and triglycerides in chowfed WT mice, and triglyceride levels in Paigen diet-fed apoE Ϫ / Ϫ mice fell after IMM-H007 treatment (supplementary Tables 1-3). Although hepatic cholesterol and triglyceride levels were not different in chow diet-fed mice treated with IMM-H007 (supplementary Fig. 3B, C), they were reduced by 200 mg/kg IMM-H007 in Paigen diet-fed mice (supplementary Fig. 3D, E).

IMM-H007 enhances cholesterol effl ux in vitro via ABCA1
To determine whether the elevated HDL caused by IMM-H007 has altered function, we performed cholesterol effl ux assays on macrophages using PEG-HDL isolated from IMM-H007-treated WT chow-fed mice. As shown in Fig. 5A, B ,

IMM-H007 suppresses atherosclerotic lesion formation in apoE ؊ / ؊ mice
We next tested the hypothesis that the enhanced RCT resulting from the suppression of ABCA1 protein degradation by IMM-H007 decreases atherosclerotic plaque formation in hypercholesterolemic mice. As expected, positive control treatment with probucol caused a decrease in lesion abundance ( Fig. 7A-E ). Compared with vehicle-treated apoE Ϫ / Ϫ mice with HDL cholesterol levels of 4.50 ± 0.56 mmol/l, IMM-H007 reduced the number and size of plaques in the aortic arch region ( Fig. 7A, B ). There were no adverse effects of IMM-H007 on liver enzymes (supplementary Fig. 7). The effect was maximal at 200 mg/kg IMM-H007 treatment, which yielded HDL levels of 7.22 ± 1.33 mmol/l (supplementary Table 3). In the entire en face aorta, the amount of lipid-laden plaque area was reduced 3-to 4-fold in mice treated with 200 mg/kg IMM-H007 ( Fig. 7B, D ). Oil Red O staining of cross-sections of the aortic root also revealed that

IMM-H007 promotes macrophage RCT in vivo
Next, we performed in vivo RCT assays to assess whether the in vitro observations above can extend to in vivo circumstances in apoE Ϫ / Ϫ mice.  4% increase, Fig. 6A ). The level of tracer was also increased in the bile, liver, and feces of IMM-H007treated mice following doses of 100 mg/kg and 200 mg/kg ( Fig. 6 B-D ). An increase in macrophage RCT to bile, liver, and feces was also demonstrable in WT mice fed with normal chow diet and treated with 200 mg/kg IMM-H007 (supplementary Fig. 5). Taken together, these observations indicate that IMM-H007 increases the initial step in cholesterol effl ux (ABCA1-mediated cholesterol effl ux) to result in the promotion of overall RCT in vivo. inhibits calpain-mediated ABCA1 degradation. Following calpain silencing, IMM-H007 had no impact on ABCA1 turnover, indicating that the primary effect of the small molecule is on calpain-mediated ABCA1 degradation.
The small molecule IMM-H007 was derived from cordycepin, which is an adenosine analog isolated from the fungus Cordyceps militaris , and synthesized as a purine analog. We previously found that cordycepin decreases LDL cholesterol levels in vivo ( 34 ), but its effi cacy is limited due to poor bioavailability . IMM-H007 was synthesized and observed to stimulate the phosphorylation of AMPK and to decrease lipid biosynthesis ( 35 ). AMPK activation by IMM-H007 may partially explain the observed declines in LDL cholesterol and triglycerides in plasma and liver because AMPK activation attenuates cholesterol synthesis, lipogenesis', and triglyceride synthesis ( 36 ). However, the observed increases in ABCA1 protein stability, in HDL cholesterol effl ux capacity, and in RCT with IMM-H007 are not likely related to AMPK activation because previous studies in macrophage-derived foam cells have shown that although AMPK stimulation increases ABCG1 expression, it has no effect on ABCA1 ( 37 ). Whether IMM-H007 binds directly to ABCA1 to impact its fate, and/or whether other processes participate now warrant investigation.
It is known that the ABCA1 inhibitor probucol reduces HDL cholesterol levels yet decreases atherosclerosis . This is thought to be due to hepatic ABCA1 antagonism and the diversion of HDL cholesterol from effl ux back into the plasma to excretion in the bile ( 25 ). The two oxidized products of probucol, spiroquinone and diphenoquinone, have been found to reduce ABCA1 protein degradation. Studies in cholesterol-fed rabbits have further shown that spiroquinone and diphenoquinone decrease lipid deposition in atherosclerotic lesions, and this was associated with increases in plasma HDL ( 10 ). However, how the inhibition IMM-H007 suppresses plaque formation in a dose-dependent manner ( Fig. 7C, E ), paralleling what was observed in the aortic arch region. In addition, IMM-H007 treatment resulted in reduced macrophage accumulation in the atherosclerotic plaque (supplementary Fig. 6). Together, these outcomes indicate that the blunting of degradation of the ABCA1 protein by IMM-H007 treatment increases both HDL quantity and function, resulting in reduced atherosclerotic plaque generation.

DISCUSSION
In the current work, we show that the small molecule IMM-H007 inhibits intracellular ABCA1 protein degradation by suppression of calpain activity and ABCA1 internalization, leading to an elevation of cell surface-resident ABCA1 in macrophages, increased circulating HDL that has enhanced capacity to effl ux cholesterol, the promotion of ABCA1-mediated RCT, and the alleviation of the formation of atherosclerosis in apoE Ϫ / Ϫ mice. Thus, via effects on both ABCA1-expressing cells and circulating HDL function, the attenuation of ABCA1 degradation by IMM-H007 results in greater capacity for HDL to mediate RCT, and this likely underlies the atheroprotection that the intervention affords.
The ABCA1 protein undergoes degradation via both calpain-mediated and ubiquitin-mediated processes ( 18,32 ). The former is dependent on the protease calpain to hydrolyze the ABCA1 protein, and the PEST-enriched motif is required ( 30 ), and ubiquitin mediates ABCA1 protein proteolysis by both lysosomal and nonlysosomal degradation pathways ( 18 ). We determined that IMM-H007 decreases intracellular calpain activity and it also decreases Ca 2+ -dependent ABCA1 turnover, indicating that the agent  of ABCA1 protein degradation affects HDL function has been unknown. Enhanced macrophage RCT is generally believed to be the crucial mechanism by which HDL has atheroprotective potential, with the most important step being the initial effl ux of cholesterol from the macrophage to lipid-poor apoA-I via the ABCA1 pathway ( 7 ). Our fi ndings in the current study reveal that ABCA1 protein stabilization by IMM-H007 increases macrophage RCT, both by effects on macrophage ABCA1 and on HDL cholesterol effl ux capacity. In vivo this results in greater cholesterol transfer to the liver, to bile, and to feces, and this likely underlies the capacity of IMM-H007 to reduce atherosclerotic plaque formation. In contrast to LXR agonists which promote fatty liver ( 38 ), we found that hepatic cholesterol and triglyceride are decreased by IMM-H007.
Thus, the small molecule enhances ABCA1 expression without causing the adverse effects on the liver that complicate LXR-targeted treatments.
In conclusion, our study shows that the small molecule IMM-H007 inhibits ABCA1 degradation by suppressing calcium-activated calpain activity, resulting in increases in circulating HDL that has enhanced capacity to effl ux cholesterol, the promotion of ABCA1-mediated RCT, and a reduction in atherosclerotic plaque formation in hypercholesterolemic mice. These fi ndings indicate that the attenuation of ABCA1 protein degradation may primarily have cardiovascular benefi t not as a result of elevating circulating HDL levels, but instead by increasing the capacity of HDL to perform RCT. Furthermore, they reveal that IMM-H007 may be a promising lead agent candidate for enhancing HDL function.