NOVEL ASSAY OF HDL FUNCTION

FIELD OF INVENTION

The present invention relates to a method for predicting or diagnosing a cardiovascular disease or a metabolic disease associated with cardiovascular disease by determining the capacity of high- density lipoprotein (HDL) to acquire surface lipids of triglyceride-rich lipoprotein (TG L) during lipolysis.

The present invention also relates to a H DL-targeting compound for use in the prevention or the treatment of a subject which has been classified as being likely to respond by the method determining the capacity of H DL to acquire surface lipids of TGRL.

BACKGROUND OF THE INVENTION

Low levels of cholesterol carried by plasma high-density lipoprotein (HDL) has an indisputable value as a predictor of elevated cardiovascular (CV) risk. As a consequence, H DL-cholesterol (H DL- C) concentration is present in all existing calculators of CV risk. However, modifications of HDL-C levels using therapeutic approaches do not necessarily modify CV risk. It has therefore been suggested that it is not HDL-C itself that is causatively related to atheroprotection but rather some cardioprotective H DL function(s), which is (are) reflected by the simple measurement of HDL-C but cannot always be reliably estimated through this assay. Hence, the concept of enhancing HDL function which can be atheroprotective was developed.

The major atheroprotective function of HDL is presently thought to involve efflux of cellular cholesterol from arterial wall cells with its subsequent transport to the liver for excretion in a process of reverse cholesterol transport (RCT). Such "HDL flux hypothesis" involves in vitro measurement of cellular cholesterol efflux from lipid-loaded macrophages as a metric of CV risk (Khera AV et al. 2011; Rohatgi A et al. 2014). Promotion of cellular cholesterol efflux and RCT in vitro represents a therapeutic corollary of this hypothesis.

However, measurements of RCT pathways in humans using labeled cholesterol reveal absence of correlation between tissue cholesterol efflux and HDL-C (Turner et al. 2012). Moreover, cholesterol efflux from peripheral cells provides only a small contribution to HDL-C levels, which does not exceed 10%. The RCT process in its present interpretation cannot therefore explain the epidemiologic association between low HDL-C levels and elevated CV risk (Smits et al. 2014). Accordingly, there remains a significant need for methods allowing determination of HDL function indicative of the presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease in a subject.

Importantly, such method can be used for the stratification of patients to select those who may benefit from HDL-targeting therapy. Indeed, several agents, including CETP inhibitors (torcetrapib, dalcetrapib, evacetrapib) and formulations of nicotinic acid (niacin) developed to increase circulating HDL-C levels, beneficially target HDL metabolism and decrease cardiovascular disease, repeatedly failed in large-scaled clinical trials (Zakiev et al. 2017).

SUMMARY OF THE INVENTION

Although, the major function of HDL is presently thought to involve efflux of cellular cholesterol, it has been suggested that HDL can also serve as an acceptor for the surface constituents generated during lipolytic process, in particular by accepting free cholesterol and phospholipid from triglyceride-rich lipoprotein (Chajek et al. 1978, Redgrave et al. 1979, Tall et al. 1979). The importance of HDL capacity to acquire the surface lipid constituent from triglyceride-rich lipoprotein for the evaluation of the risk of cardiovascular disease has never been considered.

The inventor herein proposes that the transfer of surface remnants of triglyceride-rich lipoproteins (TGRL) to HDL during their lipolysis by lipoprotein lipase (LPL) in the postprandial phase represents a main function of HDL-C and develops a method for determining presence or risk of developing a metabolic disease associated with cardiovascular disease by determining the capacity of HDL to acquire surface lipid of triglyceride-rich lipoproteins during in vitro lipolysis as a model of the postprandial phase.

Furthermore, the method can also be used for the stratification of patients with cardiovascular disease or metabolic disease associated with cardiovascular disease to select those who may benefit from HDL-targeting therapy via enhanced removal of surface lipid from TGRL during LPL- mediated lipolysis.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1: A. Dose-dependence of Dil transfer from TGRL to HDL during LPL-induced lipolysis. HDL was isolated by density ultracentrifugation and incubated with Dil-labeled TGRL (30 mg TG/dl) and LPL (5μί) at a final concentration of 2.5, 5, 10 and 20 mg protein/dL. B. Dose-dependence of Dil transfer from TGRL to apoB-depleted plasma during LPL-induced lipolysis. ApoB-containing lipoproteins were precipitated from plasma by a mixture of phosphotungstic acid with Mg2+ and apoB-depleted plasma was incubated with Dil-labeled TGRL (30 mg TG/dl) and LPL (5μί) at final dilutions of 40, 20, 10 and 5-fold. Fluorescence is expressed relative to Dil fluorescence in TGRL before incubations.

Figure 2: In vitro assay for determining the capacity of HDL (as apoB-depleted plasma) to acquire free cholesterol (in the form of BODIPY-cholesterol): free cholesterol transfer to HDL in T2D and AMI patients is significantly decreased by -38% (p=0.000001) and -25% (p=0.000272), respectively, relative to healthy normolipidemic controls. Figure 3: In vitro assay for determining the capacity of HDL (as apoB-depleted plasma) to acquire phospholipid (in the form of fluorescent Dil-PL): Phospholipid (PL) transfer to HDL in AMI patients is significantly decreased by -33% (p<0.000001) relative to healthy normolipidemic controls.

Figure 4: Relationships of overall and CV mortality with plasma HDL-C levels and HDL capacity to acquire TopF and Dil from TGRL upon LPL-induced lipolysis. Sex-adjusted CV mortality was calculated using the data from the CANHEART Study (Ko, et al., J Am Coll Cardiol 2016; 68: 2073- 2083) and Copenhagen City Heart Study (CCHS) (Madsen, et al., Eur Heart J 2017; 38: 2478-2486) for the mean HDL-C levels observed in the five populations studied (AMI, T2D, high HDL-C, extremely high HDL-C, and controls) and plotted against the HDL-C levels (A), the mean HDL capacity to acquire TopF (C) and the mean HDL capacity to acquire Dil (D) in these populations. Relationships of HDL capacity to acquire TopF and Dil with HDL-C levels is also shown (B).

Figure 5: Relationship between increase in the transfer of free cholesterol (as TopF-cholesterol) from TGRL to HDL as a result of increasing HDL concentration and initial concentration of normolipidemic VLDL in the reaction mixture.

Figure 6: Circulating HDL-C concentrations measured in high HDL-C (human Apo-I transgenic), low HDL-C (ApoA-l knock out) and control (wild-type) mouse models.

Figure 7: The capacity of HDL to acquire free cholesterol measured in high HDL-C (human Apo-I transgenic), low HDL-C (ApoA-l knock out) and control (wild-type) mouse models. HDL was obtained by precipitation of apoB-containing lipoproteins as described above.

Figure 8: Postprandial aortic accumulation of [ ³ H]-cholesterol in vivo after a gavage of high HDL-C (human Apo-I transgenic animals), low HDL-C (ApoA-l knock out) and control (wild-type) mice with 100 μθ of [ ³ H]-cholesterol administrated with olive oil (100 μΙ). The animals were euthanized 2h after the gavage, their aortas removed and specific radioactivity (per tissue weight) measured. Figure 9: Significant correlation between postprandial aortic accumulation of [ ³ H]-cholesterol in vivo measured in high HDL-C (human Apo-I transgenic), low HDL-C (ApoA-l knock out) and control (wild-type) mice, and capacity of HDL to acquire free cholesterol from TG L during LPL-mediated lipolysis in vitro evaluated in apoB-depleted plasma of these mice. DEFINITIONS

The term "patient" or "subject" means any human being or non-human mammal. Especially it is a man or woman, at any age. In a particular embodiment, the patient is an individual at risk of cardiovascular disease or metabolic disease associated with cardiovascular disease. In particular, the subject may be at increased risk of developing cardiovascular disease or a metabolic disease associated with cardiovascular disease.

An individual at risk is an individual who is considered more likely to develop a disease state such as a metabolic disease associated with cardiovascular disease or a cardiovascular disease than an individual who is not at risk. An individual "at risk" may or may not have detectable symptoms indicative of the disease condition, and may or may not have displayed detectable disease prior to the treatment methods (e.g., therapeutic intervention). "At risk" denotes that an individual has one or more so-called risk factors. An individual having one or more of these risk factors has a higher probability of developing one or more disease(s) or physiological condition(s) than an individual without these risk factor(s). These risk factors can include, but are not limited to, history of family members developing one or more diseases, related conditions, or pathologies, history of previous disease, age, sex, race, diet, presence of precursor disease, genetic (i.e., hereditary) considerations, and environmental exposure.

"Biological sample" refers to a sample from the subject. Examples of biological sample include, but are not limited to, blood, plasma, serum, saliva, lymph, ascetic fluid, cystic fluid, urine, bile, nipple exudate, synovial fluid, bronchoalveolar lavage fluid, sputum, amniotic fluid, chorionic villi, peritoneal fluid, cerebrospinal fluid, pleural fluid, pericardial fluid, semen, saliva and sweat. Preferably, suitable biological samples according to the invention include human biological matrices, urine, plasma, serum, and human lipoprotein fractions.

"Lipoproteins" are complex particles that have a central hydrophobic core of non-polar lipids, primarily cholesterol esters and triglycerides. This hydrophobic core is surrounded by a hydrophilic layer consisting of phospholipids, free cholesterol, and apolipoproteins. Plasma lipoproteins are divided into several classes based on size, lipid composition, and apolipoprotein composition (chylomicrons, chylomicrons remnants, very low-density lipoprotein (VLDL), intermediate density lipoprotein (IDL), low density lipoprotein (LDL), high density lipoprotein (HDL) and Lp(a)).

"Triglyceride-rich lipoproteins (TGRL)" are responsible for the postprandial transfer of lipids to others cells of the body. TGRL according to the present invention can comprises chylomicrons, chylomicrons remnants, VLDL and/or IDL. In the circulation, triglycerides contained in TGRL are hydrolyzed by apo-CII-dependent activation of lipoprotein lipase (LPL) present on the endothelial surface. LPL removes triglycerides from VLDL and IDL in the same way as from chylomicrons and chylomicron remnants.

"Chylomicrons" are the largest (up to 1000 nm) and least dense (>0.95 g/ml) of the lipoproteins. They contain only 1-2% protein, 85-88% triglycerides, around 8% phospholipids, around 3% cholesterol esters and around 1% cholesterol. Chylomicrons contain several types of apolipoproteins including apo-AI, II & IV, apo-B48, apo-CI, II & III, apo-E and apo-H. Chylomicrons are produced for the purpose of transporting dietary triglycerides and cholesterol absorbed by intestinal epithelia. Chylomicron assembly originates in the intestinal mucosa. Excretion into the plasma is facilitated through the lymphatic system. In the plasma, chylomicrons acquire apoA-ll and apo-E from HDL.

"Chylomicrons remnants" are metabolic products of chylomicrons from which triglycerides have been selectively removed by the lipoprotein lipase. These chylomicrons remnants carry dietary lipids in the blood and are cholesterol-rich as compared to chylomicrons. Chylomicron remnants are typically several hundred nm in size but can be as small as 30-50 nm only. They contains apo- B48 and apo-E and have lost apo-CII.

"Very Low density Lipoproteins (VLDL)" are approximately 25-90 nm in size (MW 6-27 million Da) with a mean density of around 0.98 g/ml. They contain 5-12% protein, 50-55% triglycerides, 18- 20% phospholipids, 12-15% cholesteryl esters and 8-10 % cholesterol. VLDL also contain several types of apolipoproteins including apo-BlOO, apo-CI, II & III and apo-E. VLDL also acquire apo-CII and apo-E from plasma HDL. VLDL assembly in the liver involves the early association of lipids with apo-BlOO mediated by microsomal triglyceride transfer protein while apo-BlOO is translocated to the lumen of the endothelial reticulum (ER).

"Intermediate Density Lipoproteins (IDL)" are smaller than LDL (40 nm) and more dense (~1.0 g/ml). They contain the same apolipoproteins as VLDL. They are composed of 10-12% protein, 24- 30% triglycerides, 25-27% phospholipids, 32-35% cholesteryl esters and 8-10% cholesterol. IDLs are derived from triglyceride depletion of VLDL and can be taken up by the liver. Upon further triglyceride depletion IDL become LDL, the final product of VDL metabolism.

"High Density Lipoproteins" (HDL) are the smallest of the lipoproteins (6-12.5 nm) (MW 175-500 kD) and most dense (around 1.12 g/ml). HDL contains several types of apolipoproteins including primarily apo-AI, II & IV, apo-CI, II and III, apo-D and apo-E. HDL contain approximately 35-55% protein, 3-15% triglycerides, 24-46% phospholipids, 15-30% cholesteryl esters and 2-10% cholesterol. HDL are produced as a protein-rich particle in the liver and intestine and serve as a source of apo-CI & II and Apo-E proteins. HDL particles accumulate cholesteryl esters by the esterification of cholesterol by lecithin-cholesterol acyl-transferase (LCAT). LCAT is activated by apo-AI on HDL. HDL can acquire cholesterol from cell membranes and can transfer cholesteryl esters to VLDL and LDL by cholesteryl ester transfer protein (CETP). HDL can return to the liver where cholesterol is removed in the process of reverse cholesterol transport, primarily via scavenger receptor Bl, serving as a scavenger to free cholesterol. The liver can then excrete excess cholesterol in the form of bile acids. In a normal fasting individual, HDL concentration range from 2.0 to 3.0 g/L.

DETAILED DESCRIPTION OF THE INVENTION

Lipolysis of triglyceride-rich lipoprotein (TG L), primarily chylomicron (CM) but also very low- density lipoprotein (VLDL), is a key element of lipid metabolism aimed at the release of free fatty acids from triglyceride. Lipolysis is catalysed by lipoprotein lipase (LPL) and results in the shrinkage of the hydrophobic lipoprotein core with production of smaller-sized remnant TGRL, including CM remnants (core remnants) and IDL. Excess molecules of the surface monolayer are shred from the particles in a form of surface remnants; such molecules include surface apolipoproteins, phospholipid (PL) and free cholesterol (FC). Surface remnants are then predominantly transferred to HDL.

The inventor hypothesizes that the transfer of surface remnants of triglyceride-rich lipoproteins (TGRL) to HDL during their lipolysis phase represents a major function of HDL. The inventor develops an efficient in vitro method for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease by determining the capacity of HDL to acquire surface lipid of triglyceride-rich lipoprotein.

HDL capacity to acquire surface lipids from TGRL is determined by measuring the transfer of surface lipids (also designated by lipids hereafter) from TGRL to HDL. The present invention thus relates to an in vitro method for determining the capacity of HDL from a subject to acquire surface lipid of triglyceride-rich lipoprotein (TG L) comprising the steps of: i) providing a subject biological sample comprising HDL;

ii) providing TGRL containing surface lipids;

iii) contacting said biological sample comprising HDL to said TGRL; wherein HDL is diluted to the final concentration of 2.5 to 20 mg protein/dl (for example when isolated by ultracentrifugation) and/or HDL is diluted 40- to 10-fold relative to its concentration in plasma (for example when isolated by apolipoprotein B precipitation), and TGRL is diluted to the final concentration of 5 to 100 mg triglyceride/dl;

iv) adding lipoprotein lipase enzyme at a concentration comprised between 100 and 300 units/ml to the mixture obtained at step iii) and incubating between 30 to 180 minutes;

v) determining the quantity of lipid transferred from TGRL to HDL of subject.

Any of a number of biological samples comprising HDL can be assayed using the present method. For example, the sample may be fresh blood or stored blood or blood fractions. The sample may be a blood sample expressly obtained for the assays of this invention or a blood sample obtained for another purpose which can be subsampled for use in accordance with the methods according to the invention. In a preferred embodiment, said blood sample is a fasting blood sample. In certain embodiments, the sample is derived form a cryopreserved sample. For instance, the biological sample may be whole blood. Whole blood may be obtained from the subject using standard clinical procedures. The biological sample may also be plasma, preferably fasting plasma. Plasma may be obtained from whole blood samples by centrifugation of anti-coagulated blood. The biological sample may also be serum. The sample may be pretreated as necessary by dilution in an appropriate buffer solution, concentrated if desired, or fractionated by any number of methods including but not limited to ultracentrifugation, fractionation by fast performance liquid chromatography (FPLC), or precipitation. Any of a number of standard aqueous buffer solutions, employing one of a variety of buffers, such as phosphate, Tris, or the like, at physiological to alkaline pH can be used.

In a particular embodiment, the method is performed with HDL that has been isolated or purified from a biological sample, and preferably from subject or patient plasma. The methods suitable for separating and/or purifying the different fractions of lipoprotein are well known by the person skilled in the art (see Schumaker & Puppioe, 1986). Illustrative isolation methods include, but are not limited to ultracentrifugation, PEG precipitation, heparin MnCI2 precipitation, sodium phosphotungstate precipitation, dextran sulfate precipitation, gel filtration, fast protein liquid chromatography (FPLC) and immunoaffinity capture. Protocols for these and other HDL isolation methods are readily available. Thus, for example, illustrative, but non-limiting protocols for HDL isolation by PEG precipitation, heparin MnCI2 precipitation, sodium phosphotungstate precipitation, and dextran sulfate precipitation are described by Wieve and Smith, 1985.

In a preferred embodiment, HDL are isolated by heparin/MnCI2, dextran sulfate, sodium phosphotungstate-MgCI2, or PEG precipitation, preferably by phosphotungstate-MgCI2 precipitation. In another preferred embodiment, HDL are isolated by ultracentrifugation on the appropriate density layer, knowing that density of HDL is from 1.063 to 1.210 g/mL. In a more preferred embodiment, HDL are isolated by ultracentrifugation at a density raised to 1.210 g/mL. The fractions isolated after ultracentrifugation should be further purified by dialysis.

In a particular embodiment, the method of the invention comprises an additional step i'), conducted between step i) and step ii) or step ii) and iii), consisting in isolating HDL of said subject biological sample by precipitation.

In another particular embodiment, the method of the invention comprises an additional step i'), conducted between step i) and step ii) or step ii) and iii), consisting in isolating HDL of said subject biological sample by ultracentrifugation.

In the context of the invention, the term "purified" when referring to HDL is intended to mean that the HDL represents at least 80% of lipoprotein on a mass basis of the composition comprising it. More preferably, the term "purified" indicates that lipoprotein represents by order of preference at least 85%, 90%, 92%, 95%, 97%, 98%, 99%, 100% on a mass basis of the composition.

Triglyceride rich lipoprotein (TGRL) comprises chylomicrons, chylomicrons remnants, VLDL and/or IDL. Thus, chylomicron, chylomicron remnant, VLDL and/or IDL, preferably chylomicron and VLDL or a mixture of both can be used in the method according to the present invention.

Any of a number of biological samples containing TGRL can be assayed using the present method. As described above, the sample may be fresh blood or stored blood or blood fractions. In one embodiment, said sample is normal or healthy plasma, preferably normolipidemic plasma. In another embodiment said sample is subject or patient plasma, preferably said sample is the same plasma than subject plasma comprising HDL. The methods suitable for separating and/or purifying the different fractions of lipoprotein are well known by the person skilled in the art (see Schumaker & Puppioe, 1986). Illustrative isolation methods include, but are not limited to ultracentrifugation, precipitation, and immunoaffinity capture. In a preferred embodiment TGRL are isolated by ultracentrifugation. For instance, one can use ultracentrifugation on the appropriate density layer, knowing that density of chylomicron is inferior to 0.95 g/ml, density of VLDL is from 0.95 to 1.006 g/mL and density of IDL is from 1.006 to 1.019 g/ml. In a preferred embodiment, TGRL are isolated by ultracentrifugation at a density raised to 1.019 g/ml. The fractions isolated after ultracentrifugation can be further purified by dialysis. In a particular embodiment of the method of the invention, TGRL are isolated by ultracentrifugation of a healthy normolipidemic human biological sample. In another particular embodiment TGRL are isolated from said subject biological sample, preferably said sample is the same plasma than subject plasma comprising HDL.

In another embodiment, synthetic VLDL or chylomicron can be used. The method of the invention may be conducted with TGRL containing labeled or non-labeled lipids. In a particular embodiment, TGRL are TGRL containing labeled lipids.

Contacting TGRL with HDL

The method according to the invention further comprises a contacting step of TGRL with HDL. In a particular embodiment, the contacting step comprises mixing HDL and TGRL. HDL and TGRL can be diluted in suitable buffer. Any suitable buffer may find use in embodiments herein. For example, buffers include but are not limited to phosphate buffers, HEPES, MOPS, HEPPS, and Tris-acetate, glycine, etc. HDL, for example when isolated by ultracentrifugation, can be diluted to the final concentration of 1 to 50 mg protein/dl, preferably 2.5 to 20 mg protein/dl, more preferably 2.5 to 10 mg protein/dl, more preferably 4 mg protein/dl. HDL isolated by apolipoprotein B precipitation can be diluted 40- to 10-fold relative to its concentration in plasma, more preferably 30-fold. Labeled TGRL can be diluted to the final concentration of 5 to 100 mg triglyceride/dl, preferably 15 to 60 mg triglyceride/dl, preferably 20 to 40 mg triglyceride/dl, more preferably 30 mg triglyceride/dl.

In a preferred embodiment, the ratio HDL protein:TGRL triglyceride is comprised between 1:5 and 1:10.

Adding lipoprotein lipase enzyme The inventor reports that when lipoprotein lipase (LPL) is added to TGRL and HDL mixture, LPL is capable of hydrolyzing triglycerides that allows the transfer of TGRL surface lipid to HDL. Thus, the method according to the present invention, include the addition of LPL to the mixture of HDL and TGRL before measuring labeled lipid incorporation into HDL. Lipoprotein lipase according to the invention is an enzyme capable of hydrolyzing triglycerides contained in TGRL. The catabolism of lipids results in the transfer of surface lipid from TGRL to HDL. LPL may be obtained from animals, plants or microorganisms, or produced by genetic engineering techniques. In a preferred embodiment, LPL is obtained from bovine milk (EC n°232- 669-1), from pseudomonas sp. (EC n°232-669-l) and from Burkholderia sp. (EC n° 232-669-1). Chemically modified LPL can also be used. Examples of the chemically modified enzymes include enzymes that are modified with chemically modifying groups such as a group comprising polyethylene glycol or polypropylene glycol as a main component, a group having a copolymer of polypropylene glycol and polyethylene glycol, a group comprising a water-soluble polysaccharide, a sulfopropyl group, a sulfobutyl group, a polyurethane group and a group having the chelating function. Specifically preferred is an enzyme modified with a group comprising polyethylene glycol as a main component. Examples of the water-soluble polysaccharides include dextran, pullulan and soluble starch.

The concentration of the LPL in the method according to the invention is preferably about 100 to 300 units/ml. The concentration of the LPL is preferably adapted to the concentrations of TGRL and HDL; preferably, the ratio TGRL triglyceride (mg) : LPL activity unit is comprised between 1:300 and 1:1000.

Time incubation of LPL with labeled TGRL and HDL mixture is comprised between 5 to 180 minutes. Preferably between 30 to 180 minutes, preferably 120 minutes.

Determination of the quantity of lipid transferred from TGRL to HDL of subject In a particular embodiment, HDL capacity to acquire surface lipid from TGRL is determined by measuring the difference of quantity of lipid within HDL measured before step iii) and after step iv).

In a particular embodiment, prior measuring step v), HDL can be isolated by precipitation or ultracentrifugation as described above. The present invention thus relates to an in vitro method for determining the capacity of HDL from a subject to acquire surface lipids of triglyceride-rich lipoprotein (TGRL) further comprising the steps of: ii') measuring the quantity of lipids within HDL of subject between step i) and ii) or ii) and iii), iv') measuring the quantity of lipids within HDL between step iv) and step v) and wherein the quantity of lipids transferred from TGRL to HDL of subject of step v) is determined by measuring the difference of quantity of lipid within HDL measured in step ii') and iv'). The methods suitable for quantifying the lipids within HDL are well known by the person skilled in the art. ELISA, radioimmunoassay (RIA), electrophoresis, HPLC, FACS, immuno- turbidometric assays, capillary electrophoresis, and two dimensional gel electrophoresis with or without immunodetection methods can be used as alternatives to NMR spectroscopy for measuring lipids. Numerous methods are available for determination of cholesterol concentration, e.g., gravimetric, nephelometric, turbidimetric, or photometric methods, among others. Commercially available kits for quantitative colorimetric/fluorimetric cholesterol and cholesteryl esters determination may be used. Usually, the concentrations of total and free cholesterol (esterified cholesterol being previously precipitated by, for example, digitonin) are determined, whereas the concentration of cholesteryl esters (esterified cholesterol) is calculated from the difference between these two concentrations. Enzymatic determination of cholesterol concentration is specific and sensitive. Illustrative, non-limitative, methods for determination of phospholipids concentration, include commercially available assay kits for a quantitative colorimetric/fluorimetric phospholipid determination.

In another particular embodiment, HDL capacity to acquire surface lipids from TGRL is determined by providing TGRL containing labeled lipids and measuring the transfer of labeled lipids from TGRL to HDL.

The present invention thus relates to an in vitro method for determining the capacity of HDL from a subject to acquire surface lipids of triglyceride-rich lipoprotein (TGRL) wherein said TGRL comprise labeled lipids and wherein the quantity of lipids transferred from TGRL to HDL of subject in step v) is determined by measuring the signal of labeled lipids within HDL of subject. In other terms, the present invention relates to an in vitro method for determining the capacity of HDL from a subject to acquire surface lipids of triglyceride-rich lipoprotein (TGRL) comprising the steps of: i) providing a subject biological sample comprising HDL;

ii) providing TGRL containing labeled lipids;

iii) contacting said biological sample comprising HDL to said TGRL containing labeled lipids; wherein HDL is diluted to the final concentration of 2.5 to 20 mg protein/dl (for example when isolated by ultracentrifugation) and/or HDL is diluted 40- to 10-fold relative to its concentration in plasma (for example when isolated by apolipoprotein B precipitation), and TGRL is diluted to the final concentration of 5 to 100 mg triglyceride/dl;

iv) adding lipoprotein lipase enzyme at a concentration comprised between 100 and

300 units/ml to the mixture obtained at step iii) and incubating between 30 to 180 minutes;

v) measuring signal of labeled lipid within HDL of subject.

Labeled TGRL refers to TGRL comprising labeled lipids. In a preferred embodiment, said labeled lipids are a labeled phospholipid and/or labeled free cholesterol, more preferably a labeled free cholesterol.

The labeled TGRL is obtained by contacting at least one labeled lipid with TGRL under conditions suitable for the binding of labeled lipid, preferably phospholipid or cholesterol, to TGRL to form a labeled TGRL. After being mixed, TGRL starts to absorb the labeled lipid. There is no particular limitation as to the temperature conditions and the contact time in the step. For example, the mixture of the sample and the labeled probe may be incubated at 20°C to 37°C, preferably 35°C to 37°C for 1 minute to 24 hours, preferably 1 hour to 4 hours. The mixture may be allowed to stand or may be stirred or shaken during incubation.

The term label refers to a molecule or moiety that can be detected, for example, by performing an assay known to those of skill in the art for its detection. A detectable label, accordingly, may be, for example, (i) an isotopic label (e.g., a radioactive or heavy isotope, including, but not limited to, 2H, 3H, 13C, 14C, 15N, 31P, 32P, 35S, 67Ga, 99mTc (Tc-99m), lllln, 1231, 1251, 169Yb, and 186Re), (ii) an affinity label (e.g., an antibody or antibody fragment, an epitope, a ligand or a ligand-binding agent) (iii) and enzymatic label that produce detectable agents when contacted with a substrate (e.g., a horseradish peroxidase or a luciferase); (iv) a dye, (e.g., a colored, luminescent, phosphorescent, or fluorescent molecule, such as a chemical compound or protein).

Labeled lipid can be lipophilic dye. Suitable lipophilic dye include fluorescently-tagged lipid anchors (e.g. fluorescently-labeled fatty acid analogs). An example of labeled fatty acid analog is NBD-ceramide such as l-oleoyl-2-{6-[(7-nitro-2-l,3-benzoxadiazol-4-yl)amino]hexan oyl}-s/i- glycero-3-phosphocholine. Other exemplary lipophilic dyes include, without limitation, carboxyfluorescein, BODIPY dyes, or the Alexa Fluor series. Such dyes are known by those skilled in the art and may be chosen from a group including, but not limited to lipophilic versions of fluorescent dyes including Alexa Fluor(R) 350, Alexa Fluor(R) 405, Alexa Fluor(R) 488, Alexa Fluor(R) 532, Alexa Fluor(R) 546, Alexa Fluor(R) 555, Alexa Fluor(R) 568, Alexa Fluor(R) 594, Alexa Fluor(R) 647, Alexa Fluor(R) 680, Alexa Fluor(R) 750, BODIPY(R) FL, Coumarin, Cy(R)3, Cy(R)5, Fluorescein (FITC), Oregon Green(R), Pacific Blue, Pacific Green, Pacific Orange, Tetramethylrhodamine (TRITC), Texas Red(R), DNA stains, DAPI, Propidium Iodide, SYTO(R) 9, SYTOX(R) Green, TO-PRO(R)-3, Qdot(R) probes, Qdot(R) 525, Qdot(R) 565, Qdot(R) 605, Qdot(R) 655, Qdot(R) 705, Qdot(R) 800, other lipophilic fluorescein derivatives such as carboxyfluorescein, carbocyanine derivatives such as iD (DilC18[5]), Dil (or DilC18[3]), Dil in vegetable oil, Dilinoleyl Dil, Dilinoleyl DiO, DiO (or DiOC18[3]), DiOC14(3), hydroxyethanesulfonate, DiOC16(3), DiR (DilC18[7]), DiSC2(5), DODC (DiOC2(5)), Neuro-Dil, Neuro-Dil in vegetable oil, Neuro-DiO, Neuro- DiO in vegetable oil.

In a preferred embodiment, labeled lipid is a labeled cholesterol. Labeled cholesterol is a substance in which a label binds to a part of a molecule of cholesterol. Fluorescent-labeled cholesterol and label including a fluorophore having a polar structure are known in the art. The fluorescence-labeled cholesterol including a fluorophore having a polar structure is, for example, fluorescence-labeled cholesterol including a fluorophore having a boron-dipyrromethene skeleton, such as 23-(dipyrro-metheneboron-difluoride)-24-norcholesterol), (TopFluor Cholesterol, CAS No: 878557-19-8, available from Avanti Polar Lipids, Inc.) or a benzoxadiazole skeleton such as 25-[N-[(7-nitro-2-l,3-benzoxa diazole-4-yl)methyl] amino]-27-norcholesterol (25- NBD Cholesterol, CAS No: 105539-27-3, available from Avanti Polar Lipids, Inc.) In another preferred embodiment, labeled lipid is a labeled phospholipid, preferably fluorescent phospholipid. In such a case, suitable label used can be fluorescent label-linked fatty acids include ADIFAB fatty acid indicators phospholipids with BODIPY dye-labeled acyl chains such as BODIPY glycerophospholipids, phospholipid with DPH-labeled acyl chain, phospholipids with NBD-labeled acyl chains, phospholipids with pyrene-labeled acyl chains, phospholipids with a fluorescent or biotinylated head group, LipidTOX phospholipid and neutral lipid stains.

In a preferred embodiment, fluorescent phospholipid is selected from the group consisting of : l,l'-Dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate and l-oleoyl-2-{6-[(7-nitro-2- l,3-benzoxadiazol-4-yl)amino]hexanoyl}-s/ glycero-3-phosphocholine. In another preferred embodiment, fluorescent cholesterol is 23-(dipyrro-metheneboron difluoride)-24-norcholesterol. In a particular embodiment, labeled lipid can be isotopic labeled lipid which is identical to a corresponding unlabeled lipid but for the fact that one or more atoms are replaced by an atom having an atomic mass or mass number different from the atomic mass or mass number most commonly found in nature. Examples of isotopes that can be incorporated into compounds are hydrogen, carbon, nitrogen, fluorine such as 3H, 11C, 14C and 18F.

The signal of labeled lipid within HDL is determined by a quantitative measurement of labeled lipid within HDL, more particularly by detecting signal emitted by labeled lipid. Measurement of labeled lipid can be performed using any suitable technique known in the art.

In a preferred embodiment, the measuring step comprises quantitative measurement of fluorescent lipid within HDL. Fluorescent lipid is excited by a light source and emits a signal which is detected by a fluorometric assay. The following assays may be performed using any technique employing a device comprising a fluorimeter including: flow cytometry, microscopy, optical measurement of fluorescence, and combinations thereof. In a preferred embodiment, methods herein utilize a fluorometric plate reader. These readers have, for example, a light source which is directed from above the plate and the resultant fluorescence is detected by a detector positioned either directly above the plate or at an angle above the plate.

In a particular embodiment, isotopically labeled lipid molecules can be measured by various methods such as mass spectrometry, including but not limited to gas chromatography-mass spectrometry (GC-MS), isotope-ratio mass spectrometry, GC-isotope ratio-combustion-MS, GC- isotope ratio-pyrrolysis-MS, liquid chromatography-MS, electrospray ionization-MS, matrix assisted laser desorption-time of flight-MS, Fourier-transform-ion-cyclotron-resonance-MS, and cycloidal-MS. Method for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease in a subject

In another aspect, the present invention concerns an in vitro method for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease, in a subject comprising the steps of: i) determining the quantity of lipid transferred from TG L to HDL of a subject as described above; ii) comparing said quantity with a control value.

Examples of disorders related to metabolic diseases associated with cardiovascular disease include but are not limited to stroke, ischemic stroke, transient ischemic attack, myocardial infraction, angina pectoris, inflammatory disorder, V-related diseases and/or metabolic-related diseases: type 2 diabetes, metabolic syndrome, atherosclerosis, premature atherosclerosis, hyperlipidemia, especially hypercholesterolemia, familial hypercholesterolemia, familial combined hyperlipidemia, hypoalphalipoproteinemia, coronary heart disease, coronary artery disease, acute coronary syndrome, vascular and perivascular diseases, renovascular insufficiency, critical limb ischemia, rest pain, gangrene, restenosis, rheumatoid arthritis, dyslipidemic disorders, dyslipoproteinemia, high levels of low density lipoprotein cholesterol, high levels of very low density lipoprotein cholesterol, low levels of high density lipoproteins, high levels of lipoprotein Lp(a), high levels of apolipoprotein B, familial combined hyperlipidemia (FCH), lipoprotein lipase deficiencies, such as hypertriglyceridemia and hypoalphalipoproteinemia.

Examples of cardiovascular disorders include but are not limited to: coronary artery disease (also known as coronary heart disease and ischemic heart disease), cardiomyopathy; hypertensive heart disease, heart failure, cardiac dysrhythmias, inflammatory heart disease, endocarditis, inflammatory cardiomegaly, myocarditis, valvular heart disease, cerebrovascular disease, stroke, ischemia, peripheral arterial disease, congenital heart disease (heart structure malformations existing at birth), rheumatic heart disease.

The method for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease thus comprises comparing the quantity of lipid transferred from TGRL to HDL of a subject with a control value.

The "control value" refers to a standard value of quantity of lipid transferred from TGRL to HDL in healthy normolipidemic subjects or a population of healthy normolipidemic subjects; a normolipidemic subject is a healthy non-dislipidemic subject not having a dyslipidemia as defined by Fredrickson (Fredrickson DS, Lees RS. A system for phenotyping hyperlipoproteinemia. Circulation 1965;31:321-327).

In a particular embodiment, the quantity of lipid transferred from TGRL to HDL can be determined by measuring the difference between the quantity of lipid within HDL before step iii) and after step iv). Thus, the "control value" refers to a standard value of quantity of lipid transferred from TGRL to HDL in healthy normolipidemic subjects or a population of healthy normolipidemic subjects determining by measuring the difference between the quantity of lipid within HDL before step iii) and after step iv).

In a particular embodiment, the method for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease comprises comparing the signal of labeled lipid within HDL in the patient with a control value. Thus, the "control value" refers to a standard value of signal of labeled lipid within HDL in healthy normolipidemic subjects or a population of healthy normolipidemic subjects. A decreased signal of the quantity of lipid transferred from TGRL to HDL in the subject sample compared to the control value is indicative of presence or risk of developing a metabolic disease associated with cardiovascular disease. The terms "decreased quantity" or "decreased signal" refer to a significantly lower quantity, e.g. of more than 10%, preferably more than 20%, 30%, 40%, or 50%.

According to specific embodiments, the in vitro method for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease, in a subject is such that:

- said subject is a patient with acute myocardial infarction (AMI) and said lipid is free cholesterol, preferably labeled free cholesterol, or phospholipid, preferably labeled phospholipid;

- said subject is a patient with type 2 diabetes (T2D) and said lipid is free cholesterol, preferably labeled free cholesterol;

- said subject is an individual with high HDL-C level (between 70 and 100 mg/dL) or very high HDL- C level (more than 100 mg/dL) and said lipid is free cholesterol, preferably labeled free cholesterol.

The method according to the invention can be associated with known methods for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease in a subject.

In particular, the method for determining presence or risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease in a subject according to the present invention further comprises steps of: measuring the quantity of LDLc, triglycerides, Lp(a), apoB in a biological sample of a subject and comparing the quantity of LDLc, triglycerides, Lp(a), apoB in the subject with a control value.

An increased quantity of LDLc, triglyceride, Lp(a) or ApoB in the subject biological sample compared to the control value is indicative of the presence of presence or risk of developing a metabolic disease associated with cardiovascular disease. The term "increased signal" refers to a significantly higher quantity, e.g. of more than 10%, preferably more than 20%, 30%, 40%, or 50%.

Kit

In another aspect, the invention also provides a kit comprising TGRL containing labeled lipid, lipoprotein lipase and a control sample that is preferably a sample containing HDL obtained from an healthy normolipidemic subject. In a particular embodiment, said kit further comprises a Apo-B precipitant reagent such as sodium phosphotungstate precipitation agent to isolate HDL.

Said kit can further comprises suitable buffer. For example, buffers include but are not limited to phosphate buffers, HEPES, MOPS, HEPPS, and Tris-acetate, glycine, etc. In one embodiment, the kit is for determining the capacity of HDL to acquire surface lipid from TGRL in a subject. In another embodiment, the kit is for determining the presence or the risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease.

Control sample can be sample comprising HDL from healthy normolipidemic subjects or a population of healthy normolipidemic subjects. In particular, control sample is plasma obtained from a healthy normolipidemic subject ("normolipidemic plasma"). The control sample allows to compare HDL capacity to acquire surface lipid from TGRL of a patient to a control value.

Non-limiting examples of one or more other kit components include instructions for use; vials, containers or other storage vessels containing each of the unit doses; delivery devices such as needles, catheters, syringes, tubing and the like; and/or packaging suitable for safely and conveniently storing and/or transporting the kit.

HDL-targeting compounds

The present invention also encompasses a HDL-targeting compound for use in the prevention or the treatment of a subject which has been classified as presenting or as being at risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease with the method according to the present invention.

As used herein, the term "treatment", "treat" or "treating" refers to any act intended to ameliorate the health status of patients such as therapy, prevention, prophylaxis and retardation of the disease. In certain embodiments, such term refers to the amelioration or eradication of a disease or symptoms associated with a disease. In other embodiments, this term refers to minimizing the spread or worsening of the disease resulting from the administration of one or more therapeutic agents to a subject with such a disease.

"HDL-targeting compound" refers to compounds which raise HDL levels in circulation and/or improve or normalize HDL metabolism. HDL-targeting compounds referring to compounds which raise HDL levels in circulation are also named HDL-raising compounds. HDL targeting compounds raise HDL levels in circulation and/or improve or normalize HDL metabolism by either one of the following mechanisms: CETP inhibition/modulation, HDL mimetics, ApoA-l transcriptional regulators, PPAR agonism, LXR agonism, FXR agonist, HM74 agonism (niacin receptor) thyrotropin hormone receptor agonism, inhibitors of lipases and HDL catabolism, ApoA-l inducers, compounds which provide at least one of the HDL athero-protective activities such as compounds that would increase cellular lipid efflux (cholesterol and/or phospholipids), have antioxidant and antiinflammatory activities. CETP inhibitors represent a novel class of drugs under development, which aim to improve and/or normalise lipoprotein metabolism via inhibition of cholesteryl ester transfer protein (CETP).

HDL mimetics constitute another class of perspective drugs that mimic native HDL particles and are composed of an apolipoprotein without or with a phospholipid, two major HDL components. ApoA-l transcriptional regulators are compounds specifically enhancing hepatic production of ApoA-l at a transcriptional level.

As non-limiting examples, HDL-targeting compound is selected from the group consisting of: niacin, fibrates, statins, Apo-AI mimetic peptides (e.g., APP018, D-4F, L-4F, 6F, 5A and ATI-5261), apoA-l transcriptional up-regulators (e.g., VX-208, Resverlogix), ACAT inhibitors (e.g., avasimibe; IC-976, Pfizer; MCC-147, Mitsubishi Pharma), CETP modulators (e.g., torcetrapib, evacetrapib, anacetrapib), HDL infusion compound (e.g. MDCO-216, CSL111, CSL112; CER-001) or combinations thereof.

Examples of "HDL-targeting compound" are niacin, fibrates, glitazone, dalcetrapib, anacetrapib, evacetrapib, torcetrapib, DEZ-001 (formerly known as TA- 8995, Mitsubishi Tanabe Pharma), ATH- 03 (Affris), DRL- 17822 (Dr. Reddy's), DLBS-1449 (Dexa Medica), RVX-208 (Resverlogix), MDCO-216 (The Medicines company ; former ETC-216), CSL111 (CSL Behring), CSL-112 (Cls Behring), CER-001 (Cerenis), ApoAI- Milnano (Medicine Company), ACP-501 (Medlmmune), APP018 (Bruin Pharma).

Pharmaceutical composition

In a further aspect, the present invention also provides a pharmaceutical composition comprising an HDL-targeting compound as previously defined and a pharmaceutically acceptable excipient. Preferably, the present invention also relates to the pharmaceutical composition of the invention for use in the treatment of a subject which has been classified as presenting or as being at risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease with the method for determining the capacity of HDL from a subject to acquire surface lipid of TGRL according to the present invention. It also relates to the use of a HDL-targeting compound or a pharmaceutical composition of the invention for the manufacture of a medicament for use in the treatment of a subject which has been classified as presenting or as being at risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease with the method for determining the capacity of HDL from a subject to acquire surface lipid of TG L according to the present invention.

All the embodiments described above are also contemplated in this aspect.

The pharmaceutically acceptable excipient is selected according to the route of administration and the nature of the active ingredient, e.g. a protein, a nucleic acid or a viral particle. As used herein, the term "pharmaceutically acceptable" means approved by a regulatory agency or recognized pharmacopeia such as European Pharmacopeia, for use in animals and/or humans. The term "excipient" refers to a diluent, adjuvant, carrier, or vehicle with which the therapeutic agent is administered. As is well known in the art, pharmaceutically acceptable excipients are relatively inert substances that facilitate administration of a pharmacologically effective substance and can be supplied as liquid solutions or suspensions, as emulsions, or as solid forms suitable for dissolution or suspension in liquid prior to use.

Possible pharmaceutical compositions include those suitable for oral, rectal, topical (including transdermal, buccal, sublingual, ocular instillation), intraocular (including intravitreal, intracameral, subretinal, suprachoroidal, periocular, subconjunctival) or parenteral (including subcutaneous, intramuscular, intraspinal, intravenous and intradermal) administration. For these formulations, conventional excipient can be used according to techniques well known by those skilled in the art.

Pharmaceutical compositions according to the invention may be formulated to release the active drug substantially immediately upon administration or at any predetermined time or time period after administration.

The amount of pharmaceutical composition of the invention to be administered has to be determined by standard procedure well known by those of ordinary skill in the art.

Physiological data of the patient (e.g. age, size, and weight), the routes of administration and the disease to be treated have to be taken into account to determine the appropriate dosage. The appropriate dosage of the pharmaceutical composition of the invention may also vary if it is used alone or in combination.

Method of treatment

In a further aspect, the present invention further concerns a method for treating cardiovascular disease or metabolic disease associated with cardiovascular disease in a subject comprising administering a of HDL-targeting compound wherein said subject has been classified as presenting or as being at risk of developing a cardiovascular disease or a metabolic disease associated with cardiovascular disease with the method for determining the capacity of HDL from a subject to acquire surface lipid of TGRL according to the invention.

All the embodiments described above are also contemplated in this aspect.

By a "therapeutically efficient amount" is intended an amount of the pharmaceutical composition administered to a subject that is sufficient to prevent or treat metabolic disease associated with cardiovascular disease or cardiovascular disease.

Method of in vitro assessing the activity of a compound

In another aspect, the present invention relates to a method of in vitro assessing the activity of a compound to be tested for improving the capacity of HDL from a subject to acquire surface lipid of triglyceride-rich lipoprotein (TGRL), said method comprising the steps of:

- measuring the capacity of HDL from a subject to acquire surface lipids of triglyceride-rich lipoprotein (TGRL) on a biological sample comprising HDL obtained before administration of said compound to the subject and on a biological sample comprising HDL obtained after administration of said compound to the subject; - if said capacity is significantly improved in the sample obtained after administration of said compound to the subject, then concluding that said compound is an agent able to improve the capacity of HDL to acquire surface lipid of triglyceride-rich lipoprotein (TGRL) and thus to prevent and/or treat a cardiovascular disease or a metabolic disease associated with cardiovascular disease. The examples illustrate the invention without limiting the scope. EXAMPLES:

Experimental protocol to determine lipid transfer from TGRL to HDL upon TGRL lipolysis by LPL Material and methods

1. Preparation of Dil-labeled TGRL or TopFluor ^® Cholesterol-labeled TGRL TGRL was prepared by ultracentrifugation (XL-70 Ultracentrifuge (BECKMAN) or OptimaTM MAX- TL Ul Ultracentrifuge (BECKMAN COULTER)) of normolipidemic plasma at a density raised to 1.019g/ml.

4 mg freshly isolated and dialysed TGRL-protein was added to lipoprotein-deficient serum (LPDS) at a LPDS:VLDL ratio of 1:100 by volume. The mixture was filtered with 0.8 μιτι filter and the tube was covered with aluminium foil. Dil (Sigma, MW 933.87 g/mol) or TopFluor ^® Cholesterol (TopF) (also termed BODIPY cholesterol) was added at a Dil/TGRL phospholipid or TopF/TGRL phospholipid ratio of 1:13 by mass and the mixture was incubated overnight at 37°C under gentle stirring. Labeled TGRL was separated from unbound Dil or TopF using a PD-10 Sephadex column (GE Healthcare, SephadexTM G-25M, Lot 9571451). Triglycerides (TG) concentration (CTG) was measured in purified Dil-labeled TGRL or TopF-labeled TGRL by Microplate Reader (DYNEX TECHNOLOGIES) using a commercially available kit. The volume of Dil-labeled TGRL or TopF-labeled TGRL in the reaction mixture (VTG) required to achieve the final concentration of 30mg TG/dL was calculated according to the formula VTG in TGRL= 200/(CTG in TGRL/30). Dil fluorescence in TGRL was measured at excitation and emission wavelengths of 525 nm and 570 nm respectively at TGRL-TG of 30mg/dl in a final volume of 200 μΙ (normally about 300 fluounits). TopF fluorescence in TGRL was measured at excitation and emission wavelengths of 500 nm and 525 nm respectively at TGRL-TG of 30mg/dl in a final volume of 200 μΙ.

2. Preparation of apo-B depleted HDL HDL was isolated by phosphotungstate precipitation wherein apo-B lipoproteins were precipitated with a mixture of phosphotungstic acid with Mg2+. In such a case, HDL is also named apo-B depleted plasma. Briefly, apo-B depleted plasma was prepared by adding apoB precipitant (mixture of phosphotungstic acid with Mg2+) to a reference plasma sample (normolipidemic human, EDTA; positive control) and to a plasma sample to be studied (EDTA) at a ratio of 1:10 by volume according to manufacturer's instructions which include incubation at room temperature for 10 min and centrifugation at 13 000 rpm for 30 min.

ApoB-depleted plasma is typically used at a final dilution of 30-fold v/v, which requires the volume of apoB-depleted plasma of 6.7μί to be added to the reaction mixture of 200 μΙ total volume. 3. Preparation of HDL by ultracentrifugation

HDL was prepared by ultracentrifugation (XL-70 Ultracentrifuge (BECKMAN) or OptimaTM MAX-TL Ul Ultracentrifuge (BECKMAN COULTER)) of each plasma sample at a density raised to 1.21g/ml after removal by ultracentrifugation of all lipoproteins lighter than 1.063 g/ml.

The concentration of total protein in HDL was measured by Microplate Reader (DYNEX TECHNOLOGIES) (CTP in HDL) and the volume of HDL in the reaction mixture required to achieve the final concentration of 4 mg protein/dL was calculated according to the formula VTP in HDL = 200/(CTP in HDL/4). 4. Experimental assay

50μί of Tris buffer (0.4 M, pH 8) was added to 500 μΐ Eppendorf tubes. Dil-labeled TGRL or TopF- labeled TGRL was added to the final concentration of 30 mg TG/dL followed by HDL to the final concentration of 4 mg protein/dL, or apoB-depleted plasma to the final dilution of 30-fold (6.7μί of apoB-depleted plasma).

Reference normolipidemic apoB-depleted plasma sample was added in each series of samples to be measured for data normalization.

PBS (pH 7.2) was added to final volume of 200 μΙ. 5μί of LPL enzyme (Lipoprotein Lipase from bovine milk (SIGMA, Lot SLBF4952V)) was added on the tube wall. The tube content was mixed briefly for several seconds. All tubes were incubated in an air incubator (FIRLABO) for 2 hours at 37°C. Then, the tubes were put directly on ice to stop enzymatic reaction. 20μί of the HDL-C precipitant reagent (HDL-C Precipitation Reagent) was added to each tube. Tubes were incubated for 10 min at room temperature.

1.5 ml Eppendorf tubes were prepared for each sample and plugs were removed from all of them. 500μί Eppendorf tubes with the samples were placed inside 1.5mL Eppendorf tubes without plugs. The tubes were centrifuged at 4°C for 10 min at a maximal speed (typically 13000 rpm).

0.4μιτι x 4mm filters, 1 mL syringes and corresponding needles were prepared for all samples. The supernatant was aspirated from each sample using a syringe equipped with a needle. The needle was removed and replaced with a filter. The samples were filtered into another set of 500μί Eppendorf tubes. ΙΟΟμί of each sample was transferred in a black microplate for fluorescence reading.

A standard Dil-TGRL or TopF-TGRL sample was prepared by mixing 50 μί Tris buffer (0.4 M, pH 8) with Dil-TGRL or TopF-TGRL at a final concentration of 30 mg TG/dL and with required volume of PBS in a final volume of 200 μί. ΙΟΟμί of this sample were transferred in the black microplate for fluorescence reading and the microplate was read using the Gemini fluorescence reader with an excitation and emission wavelengths of 525 and 570 nm, respectively, for Dil-TGRL and 500 and 525 nm, respectively, for TopF-TGRL. All fluorescence values measured in HDL or in apo-B-depleted plasma samples were expressed as a percentage of fluorescence of the standard Dil-TGRL or TopF-TGRL sample. Finally, the values obtained were normalized to those measured in the reference apoB-depleted plasma.

Results 1. Dose-dependence of Dil or TopF transfer from TGRL to apoB-depleted plasma during LPL- induced lipolysis

HDL was isolated by density ultracentrifugation and incubated with Dil-labeled TGRL or TopF- labeled TGRL (30 mg TG/dl) and LPL at final concentrations of 2.5, 5, 10, 20 mg protein/dl. Apo-B containing lipoproteins were precipitated from plasma by a mixture of phosphotungstic acid with Mg2+. Apo-B-depleted plasma was incubated with Dil-labeled TGRL or TopF-labeled TGRL (30 mg TG.dl) and 5 μΐ of LPL at final dilutions of 40, 20, 10 and 5-fold.

Phospholipid (PL) transfer from lipolysed TGRL to ultracentrifuged HDL (Figure 1A) or to apoB- depleted plasma (Figure IB) could be reliably observed using fluoremetrical detection (as Dil fluorescence). Similar dose-dependences were observed for cholesterol transfer (as TopF- cholesterol). The method was reproducible and could be employed for assay development.

2. Patients with acute myocardial infarction (AMI)

The new assay of HDL function was first applied to a group of patients with acute myocardial infarction (AMI), an acute form of CV disease well known to involve low plasma levels of HDL-C. The patients (n=22) were recruited immediately on their admission before initiation of any treatment for AMI. As compared to healthy, normolipidemic age-matched control subjects (n=24), AMI patients were characterized by a classical lipid profile involving significantly elevated plasma levels of triglycerides (TG; +71%, p=0.008838) and reduced levels of HDL-cholesterol (HDL-C; - 29%, p=0.000004). LDL-C concentrations did not differ between the groups, presumably reflecting statin treatment of a part of patients before the event.

Table 1. Age (years), lipid profile (mg/dl) and HDL capacity to acquire free cholesterol (FC; as TopF-cholesterol) and phospholipid (PL; as Dil-PL) upon TGRL lipolysis by LPL in patients with acute myocardial infarction (AMI) and healthy normolipidemic controls. The lipid-acquiring capacity of HDL is expressed as % of reference plasma. Means are shown for the each group followed p-values for the difference between the groups and standard deviations in each group.

The capacity of HDL to acquire both free cholesterol (FC; in the form of fluorescent TopF- cholesterol) (Figure 2) and phospholipid (PL; in the form of fluorescent Dil-PL) (Figure 3) was significantly decreased in AMI patients by -38% (p=0.000001) and -33% (p<0.000001), respectively. These decreases were more pronounced as compared to that in HDL-C levels.

No correlation of the capacity of HDL to acquire FC and PL with plasma HDL-C, LDL-C and TG levels was observed either in the whole study population or separately within the each group.

3. Patients with type 2 diabetes (T2D)

The new assay of HDL function was equally applied to a group of patients with Type 2 diabetes (T2D), a well-established rick factor for CV disease. The patients (n=36) were treatment-naive, thereby excluding potential bias from medication. As compared to healthy, normolipidemic age- matched control subjects (n=24), T2D patients were characterized by a well-known lipid profile involving significantly elevated plasma levels of triglycerides (TG; +128%, p=0.000077) and reduced levels of HDL-cholesterol (HDL-C; -24%, p=0.000048). LDL-C concentrations did not differ between the groups.

Table 2. Age (years), lipid profile (mg/dl) and HDL capacity to acquire free cholesterol (FC; as TopF-cholesterol) and phospholipid (PL; as Dil-PL) upon TGRL lipolysis by LPL in patients with Type 2 diabetes (T2D) and healthy normolipidemic controls. The lipid-acquiring capacity of HDL is expressed as % of reference plasma. Means are shown for the each group followed by p-values for the difference between the groups and standard deviations in each group.

The capacity of HDL to acquire FC was significantly decreased in T2D patients by -25% (p=0.000272) (Figure 2). This decrease was comparable to that in HDL-C levels. No correlation of the capacity of HDL to acquire FC and PL with plasma HDL-C, LDL-C and TG levels was observed either in the whole study population or separately within the each group.

4. Subjects with high level of HDL-C (70-100 mg/dl)

The assay of HDL function was also applied to a group of subjects with high level of HDL-C.

As compared to healthy, normolipidemic age-matched controls (n=10), subjects with high HDL-C levels (defined as those between 70 and 100 mg/dl; n=20) were characterised by a lipid profile involving elevated levels of HDL-C (+34%, p<0.001) in the absence of differences in LDL-C and TG concentrations (Supplement Table 3).

Controls (n=10) High HDL-C (n=20)

Age 47 ±14 56 ±13

Sex (M/F) 3/7 4/16

TC (mg/dl) 197±28 236±47*

TG (mg/dl) 62±17 84±34

LDL-C (mg/dl) 121+33 139+45

HDL-C (mg/dl) 61+15 80±7***

Supplement table 3

Despite such marked increase in HDL-C, the capacity of HDL to acquire both FC and PL did not differ between high HDL-C and normolipidemic subjects (94±33 vs 97±22% and 111±24 vs 110±13% of a reference value, respectively), thereby diverging from the HDL-C assay.

5. Subjects with extremely high level of HDL-C (>100 mg/dl)

Finally, a group of subjects with extremely high HDL-C concentrations (defined as >100 mg/dl; n=20). Controls (n=10) Extremely high HDL-C (n=20)

Age 44+13 66+9**

Sex (M/F) 4/6 3/17

TC (mg/dl) 195+30 268±55*

TG (mg/dl) 73+13 74±32

LDL-C (mg/dl) 118+35 136+53

HDL-C (mg/dl) 61+15 117±19***

Supplement table 4

Although HDL-C was elevated almost 2-fold relative to corresponding controls (n=10; Supplement Table 4), the capacity of HDL to acquire FC was diminished by -20% (81±27 vs 101±20 % of a reference value, p<0.05) in this group, while the capacity to acquire PL was elevated by +10% (119±12 vs 108±14% of a reference value, p<0.05). As a result, the latter metric was correlated with HDL-C (r=0.72, p<0.01). Similarly, HDL capacity to acquire PL was correlated with HDL-C across all the populations studied (r=0.49, p<0.001), while no such correlation was observed for the capacity to acquire FC (r=0.10, p=0.29). It is of note that no difference in the capacity of HDL to acquire surface lipids was observed between the four control groups, validating our study design.

6. Capacity to predict the presence of CV disease or CV disease risk factors.

The new assay of HDL function was further evaluated in terms of its capacity to predict the presence of CV disease, or CV disease risk factors - in other words, to discriminate between subjects without and with disease, or with risk factors for the disease.

Indeed, HDL-C levels are firmly established to represent strong, independent and negative risk factor for CV disease and to be able to distinguish between subjects without and with CV disease. Combined together with several other established CV risk factors (primarily age, sex and LDL-C levels), circulating HDL-C concentrations ensure good prediction of the presence of CV disease. A significant part of CV risk however still remains unaccounted using even the best established constellations of risk factors. Furthermore, the HDL-C assay remains empirical, without clear mechanistic relevance linking it to underlying physiological processes. The new assay of HDL function was therefore compared with the measurement of plasma HDL-C in terms of their capacity to predict the presence of AMI (an acute form of CV disease) and T2D (a CV risk factor). The predictive accuracy of the assays was evaluated by logistic regression analysis which was either unadjusted or adjusted for established CV risk factors (age, sex and plasma LDL- C and TG concentrations). a) AMI

Parameters P for Wald Odds Confiden Akaike Nagelke Likeliho Classific ROC included in prediction statistic ratio ce informati rke R ² od ratio ation area the model per 1- interval on Chi ² odds

SD criterion ratio increase (AIC)

HDL-C

(unadjuste 0.000634 11.67 0.84 0.75-0.93 44.49 0.528 23.19 12.92 0.871 d)

FC transfer

(unadjuste 0.000074 15.71 0.95 0.93-0.98 65.76 0.410 22.57 16.83 0.819 d)

PL transfer

(unadjuste 0.000069 15.85 0.93 0.89-0.96 57.58 0.526 30.74 16.67 0.881 d)

Age, sex,

50.87 0.504 21.31 9.50 0.863 LDL-C, TG

TG 1.004-

0.020191 5.39 1.02

1.05

Sex 0.030422 4.68

Age, sex,

LDL-C, TG, 42.13 0.680 32.04 22.40 0.928

HDL-C

H DL-C 0.013252 6.13 0.84 0.73-0.96

Age, sex,

LDL-C, TG, 38.92 0.725 35.26 21.25 0.946

FC transfer

FC

0.004395 8.11 0.94 0.90-0.98

transfer

TG 1.006-

0.018916 5.50 1.04

1.07

Age, sex,

LDL-C, TG, 39.83 0.713 34.35 29.75 0.940

PL transfer

PL

0.007002 7.27 0.94 0.89-0.98

transfer

TG 1.004-

0.025634 4.98 1.03

1.06

Sex 0.037465 4.33

Age, sex,

LDL-C, TG,

28.59 0.872 47.59 460.00 0.976 HDL-C, FC

transfer FC transfer 0.015421 5.87 0.91 0.85-0.98

H DL-C 0.019598 5.44 0.74 0.57-0.95

TG 1.002-

0.038808 4.27 1.04

1.08

Age, sex,

LDL-C, TG,

37.89 0.765 38.31 66.00 0.956 HDL-C, PL

transfer

PL transfer _{n Mr} . _n . _{r Λ} „„„ 0.90-

0.036045 4.39 0.94

0.996

TG 1..000022- 0.035145 4.44 1.03

1.06

Age, sex,

LDL-C, TG,

34.74 0.804 41.44 35.20 0.980 FC transfer,

PL transfer

FC transfer 0.032793 4.56 0.95 0.90-1.0

TG 0.036677 4.37 1.03 1.002-1.1

Table 3. Logistic regression analysis of the capacity to predict the presence of AMI using measurements of plasma HDL-C vs. HDL capacity to acquire lipids upon TGRL lipolysis by LPL.

Within each groups of models, the best model is shown in bold.

In unadjusted logistic analysis, HDL-C was significantly predictive of AMI (p=0.000634) with odds ratio of 0.84 (95% confidence interval, 0.75-0.93) for a 1-SD increase in HDL-C. The capacity of HDL to acquire both FC and PL was also predictive significantly of AMI (p=0.000074 and 000069, respectively) in unadjusted analysis, with odds ratio of 0.95 (95% confidence interval, 0.93-0.98) and 0.93 (95% confidence interval, 0.89-0.96), respectively; these associations were stronger than that obtained for HDL-C (Table 3). The unadjusted logistic regression models involving HDL-C, HDL capacity to acquire FC and HDL capacity to acquire PL provided areas under the receiver- operating-characteristic curve (ROC) of 0.871, 0.819 and 0.881, respectively.

An adjusted model including only established CV risk factors (age, sex, LDL-C and TG) without HDL-C displayed comparable performance (ROC area, 0.863) which was improved by adding HDL- C to the model (ROC area, 0.928). When HDL-C was replaced by HDL capacity to acquire FC or PL, the model performance was further enhanced (ROC area, 0.946 and 0.940, respectively). When HDL-C was combined with HDL capacity to acquire FC or PL, the model performance was even better (ROC area, 0.976 and 0.956, respectively) (Table 3).

In all adjusted models studied, the HDL lipid transfer function, primarily that involving FC transfer, was a more accurate predictor of AMI as compared to HDL-C. As a corollary, the best predictive accuracy of all the models was provided by that involving traditional risk factors combined with HDL capacity to acquire FC and HDL capacity to acquire PL (ROC area, 0.980). b) T2D

When HDL-C levels and H DL lipid transfer function were analyzed for their capacity to predict T2D, qualitatively similar associations relative to those described above for AMI were observed which were however slightly less pronounced.

Parameters P for Wald Odds Confide Akaike Nagel Likeliho Classific ROC included in predictio statistic ratio nee information kerke od ratio ation area the model n per 1- interval criterion R ² Chi ² odds

SD (AIC) ratio

increase

HDL-C

0.85-

(unadjuste 0.000878 11.07 0.90 67.62 0.331 16.85 13.33 0.809

0.96

d)

FC transfer

0.95-

(unadjuste 0.000986 10.85 0.97 90.43 0.226 13.38 7.00 0.746

0.99

d)

PL transfer

(unadjuste n.s 102.09 0.032 1.72 1.57 0.572 d)

Age, sex,

57.39 0.577 33.37 25.00 0.913 LDL-C, TG

TG 1.02-

0.000941 10.94 1.04

1.07

Age, sex,

LDL-C, TG, 56.57 0.612 36.19 20.71 0.912

HDL-C

TG 1.009-

0.007890 7.06 1.03

1.06

Age, sex,

LDL-C, TG, 51.60 0.671 41.16 55.00 0.928

FC tranfer

FC 0.93-

0.000563 11.90 0.96

transfer 0.99

TG 1.02-

0.13138 6.15 1.05

1.07

Age, sex,

LDL-C, TG, 57.36 0.602 35.40 43.40 0.916

PL transfer

TG 1.02-

0.000980 10.87 1.04

1.07

Age, sex,

LDL-C, TG,

50.22 0.708 44.53 77.00 0.940 HDL-C, FC

transfer

FC 0.92-

0.013153 6.15 0.96

transfer 0.99

TG 1.01-

0.004495 8.07 1.04

1.07

Age, sex,

LDL-C, TG, 56.08 0.642 38.68 25.00 0.923

HDL-C, PL transfer

TG 1.008- 0.009475 6.73 1.03

1.06

Age, sex,

LDL-C, TG,

52.34 0.685 42.41 77.00 0.931 FC trabsfer,

PL transfer

FC 0.93-

0.017415 5.65 0.96

transfer 0.99

TG 1.02- 0.00960 10.90 1.05

1.08

Table 4. Logistic regression analysis of the capacity to predict the presence of T2D using measurements of plasma HDL-C vs. HDL capacity to acquire lipids upon TGRL lipolysis by LPL.

Within each groups of models, the best model is shown in bold.

In unadjusted logistic analysis, H DL-C significantly discriminated T2D patients from controls (p=0.000878) with odds ratio of 0.90 (95% confidence interval, 0.85-0.96) for a 1-SD increase in HDL-C. The capacity of HDL to acquire FC also discriminated significantly T2D patients and controls in unadjusted analysis (p=0.000986), with odds ratio of 0.97 (95% confidence interval, 0.95-0.99); (Table 4). The unadjusted logistic regression models involving H DL-C, HDL capacity to acquire FC and HDL capacity to acquire PL provided ROC areas of 0.809, 0.746 and 0.572, respectively. An adjusted model including only established CV risk factors (age, sex, LDL-C and TG) without HDL-C displayed a better performance (ROC area, 0.913) which was not modified by adding H DL-C to the model (ROC area, 0.912). When HDL-C was replaced by H DL capacity to acquire FC, but not PL, the model performance was further enhanced (ROC area, 0.928). When HDL-C was combined with HDL capacity to acquire FC, the model performance was even better (ROC area, 0.940) (Table 4)

In all adjusted models studied, HDL capacity to acquire FC was a more accurate predictor of T2D as compared to HDL-C, which was not significantly predictive of the disease after multiple adjustment.

7. Evaluation on the basis of CV mortality data Further evaluation of the relationship of the assay with CV disease was performed on the basis of published mortality data. When overall and CV mortality data obtained from recently published large-scale CAN HEART HDL study (Ko, D.T., et al. High-Density Lipoprotein Cholesterol and Cause- Specific Mortality in Individuals Without Previous Cardiovascular Conditions: The CANH EART Study. J Am Coll Cardiol 68, 2073-2083 (2016)) and Copenhagen City Heart Study (Madsen, CM., Varbo, A. & Nordestgaard, B.G. Extreme high high-density lipoprotein cholesterol is paradoxically associated with high mortality in men and women: two prospective cohort studies. Eur Heart J 38, 2478-2486 (2017)) were recalculated for the mean HDL-C levels observed in the five populations studied by us (AMI, T2D, subjects with high and extremely high HDL-C, and pooled controls) and plotted against HDL-C, a curvilinear non-significant U-shaped relationship was observed as reported (Fig. 4, A). A similar, but mirror-like, U-shaped relationship was observed between HDL capacity to acquire TopF, but not Dil, and HDL-C (Fig. 4, B). As a result, when HDL-C was replaced by the mean values of HDL capacity to acquire FC, linear relationships with both overall and CV mortality was obtained (r ² from 0.79 to 0.97, p from 0.043 to 0.002), indicative of the significant relationship of mortality with the assay outcome (Fig. 4, C). The divergence between HDL-C and HDL capacity to acquire FC reflected only weak modifications in the both mortality and the assay outcome in subjects with high HDL-C as well as elevated mortality paralleled by reduced FC transfer to HDL at extremely high HDL-C. By contrast, no significant association between the mortality and HDL capacity to acquire PL was observed (Fig. 4, D).

8. Stratification of patients Potential use of the method for the stratification of patients to select those who may benefit from HDL-targeting therapy is based on non-linear dose-dependences between free cholesterol transfer from TG L to HDL upon LPL-mediated lipolysis and HDL concentration in the reaction mixture (Figure 1). This figure shows that increasing HDL concentration enhances the transfer of free cholesterol only at low to moderate concentrations of HDL. By contrast, the transfer is not accelerated and can even be diminished when HDL concentration is further increased to high levels.

Additional measurements of the dependence of the cholesterol transfer on HDL concentration reveal that such dependence differs as a function of VLDL concentration in the reaction mixture (Figure 5). Specifically, increasing HDL-C concentration enhances the transfer of free cholesterol only at low concentrations of normolipidemic VLDL. Such situation is characterized by low lipolytic fluxes in the reaction mixture and corresponds to suboptimal lipolysis in vivo which may occur under conditions of low LPL activity and/or high levels of apoCIII, which result in hypertryglyceridemia. Furthermore, the figure shows that small to moderate increases in HDL concentrations better impact the transfer of free cholesterol as compared to large increases in HDL.

By contrast, influence of HDL concentration on the free cholesterol transfer is weakened at high concentrations of normolipidemic VLDL, which generate high lipolytic fluxes in the reaction mixture and correspond to optimal lipolysis in vivo as it occurs under conditions of normal or high LPL activity and/or low levels of apoCIII, which result in normotryglyceridemia. Only small increases n HDL concentration appear to beneficially impact free cholesterol transfer from TG L to HDL under these conditions, whereas larger increases in HDL concentration do not accelerate the transfer. Together, these data provide an in vitro basis for the selection of patients as a function of their triglyceride and HDL concentrations in plasma for an efficient HDL-targeting therapy. Furthermore, the data may allow tailoring HDL concentration increase as a function of patient lipid phenotype, thereby deepening personal medicine approach.

9. Murine Models a) Assessment of HDL function in mouse models of dyslipidemia and atherosclerosis

Pathophysiological relevance of the assay was further evaluated in vivo using murine models of lipid metabolism and atherosclerosis.

Three mouse strains displaying marked differences in circulating levels of HDL-C were selected for these studies, specifically wild-type mice, dyslipidemic mice with a disrupted gene of apolipoprotein A-l (apoA-l), the major HDL protein, featuring low HDL levels and accelerated atherosclerosis and hyperlipidemic mice transgenic for human apoA-l featuring high HDL levels.

ApoA-l knock-out, human apoA-l transgenic and wild-type mice were purchased from Jackson Labs. Blood was obtained from the tail vein, mixed with EDTA and centrifuged to obtain plasma. HDL was isolated from the plasma by apoB precipitation using a commercially available reagent and the capacity of HDL to acquire PL and FC was measured in vitro as described above.

As expected, HDL-C concentrations were reduced by -82% in apoA-l knock-out mice and were elevated by +72% in apoA-l transgenic animals (Figure 6). Consistent with these data, the capacity of HDL to acquire FC measured upon TGRL lipolysis by LPL in vitro was reduced by -22% in apoA-l knock-out mice and elevated by +47% in apoA-l transgenic animals (Figure 7). b) In vivo assay to evaluate the capacity of HDL to acquire surface fragment of TGRL during lipolysis in mice

Mice were given 100 μθ of 3H-labelled cholesterol mixed with 100 μΙ olive oil by oral gavage. Two hours later, mice were euthanized, organs and blood removed and their mass and radioactivity determined. HDL was isolated from plasma by apoB precipitation as described above. When all the mice received an oral gavage of 3H-cholesterol administrated together with 100 μΙ of olive oil, apoA-l knock-out mice accumulated significantly less radioactivity (-68%) in the HDL fraction 2h after the gavage as compared to wild-type animals; no significant difference was observed between wild-type and apoA-l transgenic mice (Figure 8). Mirroring these differences, postprandial accumulation of 3H-cholesterol in the aorta was significantly (6.1-fold) enhanced 2h after the gavage in apoA-l knock-out mice relative to wild-type animals, while no difference between wild-type and apoA-l transgenic mice was found. As a consequence, the capacity of HDL to acquire free cholesterol (FC) in vitro was significantly and negatively correlated with the aortic accumulation of 3Hcholesterol in all mice, attesting the former as a negative metric of postprandial atherosclerosis (Figure 9).

Conclusion: Together, these data provide strong and unequivocal evidence that HDL capacity to acquire primarily FC but also phospholipids (PL) upon TG L lipolysis by LPL in vitro, measured using the present new assay as a transfer of fluorescent TopF-cholesterol or Dil-PL, represents a better biomarker of CV risk as compared to plasma concentrations of HDL-C presently employed in clinical practice. These results also demonstrate that our assay provides correct evaluation of physiologically relevant HDL function, whereas measurement of HDL-C is empirical and not directly reflective of physiologic pathways.

Finally, the data provide an in vitro evidence for the suitability of the method for the stratification of patients with metabolic disease associated with cardiovascular disease, or with cardiovascular disease itself, to select those who may specifically benefit from HDL-targeting therapy.

REFERENCES

Chajek T, Eisenberg S. Very low density lipoprotein. Metabolism of phospholipids, cholesterol, and apolipoprotein C in the isolated perfused rat heart. J Clin Invest. 1978 Jun;61(6):1654-65.

Khera AV, Cuchel M, de la Llera-Moya M, Rodrigues A, Burke MF, Jafri K, French BC, Phillips JA, Mucksavage ML, Wilensky RL, Mohler ER, Rothblat GH, Rader DJ. Cholesterol efflux capacity, high- density lipoprotein function, and atherosclerosis. N Engl J Med. 2011 Jan 13;364(2):127-35Rohatgi A et al. HDL cholesterol efflux capacity and incident cardiovascular events. N Engl J Med. 2014 Dec 18;371(25):2383-93.

Ko DT, Alter DA, Guo H, Koh M, Lau G, Austin PC, Booth GL, Hogg W, Jackevicius CA, Lee DS, Wijeysundera HC, Wilkins JT, Tu JV. High-Density Lipoprotein Cholesterol and Cause-Specific Mortality in Individuals Without Previous Cardiovascular Conditions: The CANHEART Study. J Am Coll Cardiol. 2016 Nov 8;68(19):2073-2083 Madsen, CM., Varbo, A. & Nordestgaard, B.G. Extreme high high-density lipoprotein cholesterol is paradoxically associated with high mortality in men and women: two prospective cohort studies. Eur Heart J 38, 2478-2486 (2017)

Redgrave TG, Small DM. Quantitation of the transfer of surface phospholipid of chylomicrons to the high density lipoprotein fraction during the catabolism of chylomicrons in the rat. J Clin Invest. 1979 Jul;64(l):162-71.

Schumaker VN, Puppione DL. Sequential flotation ultracentrifugation. Methods Enzymol. 1986;128:155-70.

Smits LP, Kootte RS, Stroes ES. Reversal of atherosclerosis with apolipoprotein Al: back to basics. Atherosclerosis. 2014 Jan;232(l):217-9.

Tall AR, Robinson LA. Absence of liquid crystalline transitions of cholesterol esters in reconstituted low density lipoproteins. FEBS Lett. 1979 Nov l;107(l):222-6.

Turner S, Voogt J, Davidson M, Glass A, Killion S, Decaris J, Mohammed H, Minehira K, Boban D, Murphy E, Luchoomun J, Awada M, Neese R, Hellerstein M. Measurement of reverse cholesterol transport pathways in humans: in vivo rates of free cholesterol efflux, esterification, and excretion. J Am Heart Assoc. 2012 Aug;l(4):e001826.

Wiebe DA, Smith SJ. Six methods for isolating high-density lipoprotein compared, with use of the reference method for quantifying cholesterol in serum. Clin Chem. 1985 May;31(5):746-50.

Zakiev E, Feng M, Sukhorukov V, Kontush A. HDL-Targeting Therapeutics: Past, Present and Future. Curr Pharm Des. 2017;23(8):1207-1215