Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Dec 1.
Published in final edited form as: J Diabetes Complications. 2020 Jul 31;34(12):107693. doi: 10.1016/j.jdiacomp.2020.107693

Glycation of HDL blunts its anti-inflammatory and cholesterol efflux capacities in vitro, but has no effect in poorly controlled type 1 diabetes subjects

Diego Gomes Kjerulf 1, Shari Wang 1, Mohamed Omer 1, Asha Pathak 1, Savitha Subramanian 1, Chang Yeop Han 1, Chongren Tang 1, Laura J den Hartigh 1, Baohai Shao 1, Alan Chait 1
PMCID: PMC7669727  NIHMSID: NIHMS1616813  PMID: 32900591

Abstract

Background:

High-density lipoproteins (HDL) modified by glycation have been reported to be dysfunctional. Little is known regarding the anti-inflammatory effects on adipocytes of glycated HDL.

Aims:

We tested whether modification of HDL in vitro by glycolaldehyde (GAD), malondialdehyde (MDA) or glucose affected HDL’s anti-inflammatory properties and ability to promote cholesterol efflux. To determine whether similar changes occur in vivo, we examined modifications of apolipoprotein A1 (APOA1) and APOA2 and anti-inflammatory and cholesterol efflux properties of HDL isolated from subjects with type 1 diabetes in poor glycemic control.

Results:

In vitro modification with both GAD and MDA blunted HDL’s ability to inhibit palmitate-induced inflammation and cholesterol efflux in adipocytes. Modification of HDL by glucose had little impact on HDL function, like the response using HDL isolated from subjects with diabetes. Mass spectrophotometric analysis revealed that lysine residues in APOA1 and APOA2 of HDL modified by GAD and MDA in vitro differed from those modified by glucose, which resembled that seen with HDL from patients with type1 diabetes.

Conclusions:

Modification of lysine residues in HDL by GAD and MDA in vitro does not mirror the HDL glycation in vivo in patients with diabetes, but resembles HDL modified in vitro by glucose.

Keywords: glycation, inflammation, cholesterol efflux, adipocytes, macrophages

1. Introduction

The World Health Organization estimates that nearly 10% of the adult population worldwide is living with diabetes mellitus (DM), placing DM as a significant global disease. Hyperglycemia is a hallmark of DM and triggers cellular responses that result in increased oxidative stress and inflammation (Newsholme, Cruzat, Keane, Carlessi, & de Bittencourt, 2016), and is believed to underly important cellular mechanisms leading to microvascular complications of DM. Hyperglycemia also has been implicated in the pathogenesis of macrovascular disease in DM (Wong, Nicholls, Tan, & Bursill, 2018).

One potential mechanism by which hyperglycemia might cause diabetic complications is increased formation of advanced glycation-end products (AGE) or advanced lipid peroxidation end-products (ALE) (Brownlee, 2001). AGEs are commonly formed by the initial non-enzymatic reaction between the free amino group of lysine or arginine residues in proteins with reducing sugars, including glucose, which generates early-glycosylated compounds, termed Amadori products. After several days and further irreversible rearrangements, AGEs are formed. Moreover, highly reactive carbonyl compounds such as malondialdehyde (MDA), derived from the oxidation of polyunsaturated fatty acids, can react with lysine residues in proteins to generate modified products called ALEs. The formation of AGEs and ALEs can be accelerated by glycolaldehyde (GAD), lipid and amino acid degradation, autoxidation of monosaccharides, and fragmentation of Amadori product intermediates (Ott et al., 2014). Methylglyoxal, a byproduct of glycolysis, is another AGE precursor.

Association of DM-related complications with glycation of plasma proteins, including those in high-density lipoproteins (HDL), has been suggested by others (Ahmad et al., 2017; Curtiss & Witztum, 1985; Gomes et al., 2016; Kamtchueng Simo, Ikhlef, Berrougui, & Khalil, 2017). Glycation products have been found in several proteins present in HDL, and glycation or oxidation of HDL’s major apolipoprotein, APOA1, may result in a loss of the ability of HDL to facilitate cholesterol efflux in DM (Pu et al., 2013). HDL reconstituted with in vivo glycated-APOA1 showed a loss of its anti-inflammatory properties in endothelial cells, which was attributed to impaired intracellular signaling and pro-oxidative changes (Nobecourt et al., 2010). HDL isolated from subjects with type 2 DM (T2DM) demonstrated less efficient cholesterol efflux from macrophages than HDL isolated from healthy individuals (Kashyap et al., 2018). Additionally, the presence of the metabolic syndrome in T2DM, characterized in part by hyperglycemia, obesity and a chronic inflammatory status, potentially alters HDL protein composition, a feature pivotal to the efficiency of HDL’s anti-inflammatory properties (Wong et al., 2018).

Previous studies from our group have shown that HDL isolated from normal mice and healthy human subjects inhibits the expression of pro-inflammatory genes in hypertrophic adipocytes in which inflammation was stimulated by exposure to palmitate (PA) (Han et al., 2016). However, HDL from inflamed mice and human subjects with evidence of inflammation failed to inhibit PA-mediated inflammation, which was attributed in part to the presence of serum amyloid A (SAA) in HDL from these inflamed subjects (Han et al., 2016). To specifically determine whether glycation of HDL, as distinct from potential changes in HDL’s structure and function that can be associated with T2DM, the metabolic syndrome and chronic low grade inflammation, affects these anti-inflammatory properties of HDL, we tested the ability of HDL glycated in vitro and isolated from subjects with type 1 DM (T1DM) in poor glycemic control to inhibit the inflammatory response elicited by PA in cultured adipocytes. We specifically chose to study T1DM patients to minimize the chronic inflammation and high SAA levels (Han et al., 2016) seen in T2DM and the metabolic syndrome (Griffiths et al., 2017; Kumon, Suehiro, Itahara, Ikeda, & Hashimoto, 1994; Reddy, Lent-Schochet, Ramakrishnan, McLaughlin, & Jialal, 2019). The in vivo component of our study was solely to determine whether the changes in HDL structure and function that we observed in vitro had an in vivo counterpart. To this end, we chose to study T1DM subjects in poor glycemic control to maximize the chance of emulating in vitro glycation and to minimize the effects of inflammation. In addition to evaluating the effect of these in vitro and in vivo generated HDL samples on the ability of HDL to inhibit palmitate-induced inflammation in adipocytes, cholesterol efflux capacity in both adipocytes and macrophages and post-translational modification of the major HDL apolipoproteins were analyzed.

2. Methods

2.1. Isolation of HDL and in vitro glycation

Healthy human subjects (n=9) were recruited from the University of Washington South Lake Union campus. Poorly controlled type 1 diabetic subjects (HbA1c > 9.8 ± 0.2%, n=12) were recruited from the Diabetes Care Center at the University of Washington Medical Center (Table 1). After obtaining informed written consent for blood collection, HDL (d = 1.063–1.210 g/ml) was isolated by sequential ultracentrifugation from the plasma of normal healthy human volunteers who had fasted overnight (ND-HDL) and from the plasma of non-fasted, poorly controlled diabetic subjects (DM-HDL) as previously described (Han et al., 2016). Protein concentration was determined using a Pierce BCA protein kit, with bovine gamma globulin as the standard.

Table 1: Metabolic variables in control subjects and subjects with type 1 diabetes.

Values shown are means ± SEM.

Control (n=9) Diabetic (n=12) p
Glucose (mg/dL) 89.4 ± 1.7 206 ± 22.7 0.0003
Cholesterol (mg/dL) 171.3 ± 6.3 184 ± 11.4 0.35
Triglycerides (mg/dL) 75.4 ± 7.4 116.7 ± 19.1 0.09
HDL-C (mg/dL) 65.2 ± 54.5 ± 5.9 0.93
LDL-C (mg/dL) 91.1 ± 6.4 97 ± 8.3 0.60
VLDL-C (mg/dL) 15 ± 1.4 23.4 ± 3.8 0.08
Open in a new tab

GAD-modified HDL was prepared by incubating 0.8 mg/mL of HDL with 0.5 mM (GAD0.5-HDL) or 1 mM (GAD1-HDL) of GAD (Sigma Chem. Co, St Louis, USA) in 50 mM of phosphate buffer solution containing 100 μM of DTPA for 24 hours. Similarly, using the same buffer solution, HDL was modified by incubation with 0.1 mM (MDA0.1-HDL) or 0.2 mM (MDA0.2-HDL) of MDA for 24 hours. MDA was prepared immediately before use by rapid acid hydrolysis of maloncarbonyl bis-(dimethylacetal) (Sigma Chem Co, St Louis, USA). Briefly, 20 μl of 1 M HCl was mixed with 200 μl of maloncarbonyl bis-(dimethylacetal), and the mixture was incubated at room temperature for 45 min. The reaction mixture was diluted with 980 μl of 50 mM phosphate buffer (pH 7.4). After further dilution, the MDA concentration of the stock solution was determined by absorbance at 266 nm, using ϵ = 13,700 m−1 cm−1 (Shao et al.). HDL modified by glucose (G) (Sigma Chem. Co, St Louis, USA) was facilitated by incubation with either 5 mM (GLU5-HDL), 10 mM (GLU10-HDL), or 25 mM (GLU25-HDL) in DTPA buffer or PBS alone [control pooled HDL (CON-HDL)] for 4 days at 37 °C in a water bath under sterile conditions and nitrogen atmosphere in the dark. CON-HDL, GLU-HDL, GAD-HDL and MDA-HDL were dialyzed against PBS for 24 hours, with buffer changes every 8 hours. Protein concentrations of the dialyzed samples were determined, after which the samples were filtered and maintained in small aliquots at 4°C for no longer than 2 weeks.

2.2. Inhibition of palmitate-induced inflammation in adipocytes

3T3-L1 pre-adipocytes obtained from the American Type Tissue Culture Collection were propagated and differentiated, as described previously (Han et al., 2016). Fully differentiated adipocytes were pretreated with the various types of in-vitro modified HDL (10, 25, or 50 μg protein/mL) for 6 hours, and then washed 3 times with DMEM. Adipocytes then were incubated for 24 hours with 250 μmol/L of BSA-conjugated PA (3:1 molar ratio) for measurement of gene expression by real-time PCR (RT-PCR). RT-PCR was performed using the TaqMan Master kit (Applied Biosystems) in the ABI prism 7900HT system. Saa3, Ccl2 and Gapdh primers with FAM probes were obtained from ThermoFisher. Each sample was analyzed in triplicate and normalized using Gapdh as control. Some samples also were normalized with a second housekeeping gene, β−2-microglobulin (B2m) and showed similar results. Suppression of gene expression by pre-exposure to the various HDL preparations was used as a measure of the anti-inflammatory property of HDL.

2.3. Cholesterol efflux from adipocytes

To measure the cholesterol efflux capacity of in vitro glycated HDL, control (CON-HDL), HDL from non-diabetic control (ND-HDL) and subjects with T1DM (DM-HDL), fully differentiated 3T3-L1 adipocytes were radiolabeled by incubating the cells with 1 μCi/mL 3H-cholesterol (PerkinElmer) in DMEM containing 1 mg/mL fatty acid–free BSA overnight. Cells were then washed and incubated with DMEM/fatty acid–free BSA without or with 50 μg protein/mL of the various HDL preparations for 6 hours at 37°C. Medium was collected and filtered, and the 3H-cholesterol content of medium and cells was quantified. The fraction of total 3H-cholesterol released into the medium was calculated as (counts in the medium with HDL – counts in the medium without HDL)/(counts in the medium + counts in the cells) × 100.

2.4. Cholesterol efflux from macrophages

The cholesterol efflux capacity of the various types of HDL also was measured in J774 macrophages, as described previously (de la Llera-Moya et al., 2010). Briefly, after confluency cells were washed with PBS and radiolabeled with 0.5 μCi/mL 3H-cholesterol in DMEM/BSA(1mg/mL) +L-glutamine. The lipid transporter ABCA1 was induced by 8CTPcAMP (Sigma). After 24 hours, cells were washed with PBS and cultured in DMEM/BSA. Cholesterol efflux was then assessed by incubation with 0.4 μg of the various HDL preparations for 4 hours, followed by collection and filtration of media in 96-well filter plates (AcroPrep Advance #8039). Samples were then centrifuged at 1500g for 3 min, rinsed with 20 μL dH20 and recentrifuged for 1 min. Cell were removed from plate with 250 μL of 0.2N of NaOH for 2 hours in room temperature. ABCA1-dependent cholesterol efflux was calculated as the percentage of total 3H-cholesterol (medium plus cell) released into the medium by mifepristone-treated J774, after subtraction of the value obtained with J774 cells not expressing ABCA1.

2.5. Digestion of HDL for mass spectrometry

Following the addition of freshly prepared methionine (5 mM final concentration) in 20% acetonitrile and 100 mM NH4HCO3, 5 μg of HDL proteins were reduced with dithiothreitol and then alkylated with iodoacetamide. After adding 0.2 μg of isotope-labeled [15N]-APOA1 (as the internal standard), the HDL was incubated at 37 °C with 20:1 (w/w, proteins/enzyme) of sequencing grade modified trypsin overnight. Digestion was halted by acidifying the reaction mixture (pH 2 to 3) with trifluoroacetic acid, and the samples were dried and stored at −80 °C until mass spectrometric (MS) analysis.

2.6. Liquid chromatography-electrospray ionization tandem mass spectrometric (LC-ESI-MS/MS) analysis of glycated APOA1 and APOA2 in HDL

To quantitatively measure the relative levels of glycated APOA1 and APOA2 in HDL, we used a targeted proteomics with isotype-dilution parallel-reaction monitoring (PRM) as previously reported (Kanter et al., 2019). Briefly, LC-ESI-MS/MS analyses were performed in the positive ion mode with an ultrahigh-resolution accurate mass Orbitrap Fusion Lumos Tribrid Mass Spectrometer (Thermo Fisher Scientific, San Jose, CA) coupled to a nanoACQUITY UPLC (Waters, Milford, MA). A linear gradient of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B) was used for the separation. After the dried peptide digests were reconstituted in 0.1% formic acid, HDL peptide digests (equivalent to 0.2 μg of proteins) were desalted on a C-18 trap column (0.1 × 40 mm) packed in house with Magic C-18 reverse-phase resin (5 μm; 100 Å; Michrom Bioresources) for 7 min at 3 μL/min in 99% solvent A, and then separated using a C-18 analytical column (0.1 × 200 mm) packed in house with Magic C-18 reverse-phase resin (5 μm; 100 Å; Michrom Bioresources). The column was kept at room temperature and the peptides were eluted from the trap column onto the analytical column at a flow rate of 0.5 μL/min and separated using a multistep gradient as follows: 1% to 7% solvent B in 1 min; 7% to 25% solvent B in 24 min; and 25% to 35% solvent B in 6 min; 35% to 80% solvent B in 5 min. The column was subsequently washed for 3 min at 80% B and re-equilibrated at 99% A for 12 min. The mass spectrometer was operated in data-independent acquisition PRM mode as previously described (Kanter et al., 2019).

2.7. Quantification of glycated APOA1 and APOA2 in HDL

Product yields of different glycation modified peptides in APOA1 or APOA2 of HDL were quantified with reconstructed ion chromatograms of precursor and product peptides in the same sample (Shao & Heinecke, 2008). Product yields of modified peptides were calculated from the ratio of the peak area of product peptide of APOA1 relative to the sum of precursor and product peptides. Thus, product yield (mmol/mol) = 1000 × peak area of product peptide/(peak area of corresponding precursor peptide + peak area of product peptide). When the product peptide was cross-linked by MDA, the product ion peak area was compared with the sum of the peak area of cross-linked peptide and the peak areas of all precursor peptides that formed the cross-linked product peptide. Thus, product yield (mmol/mol) = 1000 × (peak area of cross-linked product peptide/(peak area of cross-linked product peptide + sum of peak areas of all precursor peptides). This method assumes that the precursor ions and the product ions have similar MS response characteristics (Shao & Heinecke, 2008).

2.8. Statistical Analyses

Statistical analyses were done using GraphPad Prism 7. Each type of HDL was compared by one-way ANOVA followed by Bonferroni/Dunn post-hoc analysis. Comparisons between ND-HDL and DM-HDL were made by t-test followed by Bonferroni/Dunn post-hoc analysis. Data are expressed as mean ± SEM, unless otherwise indicated. A p-value of <0.05 was considered significant.

3. Results

3.1. Suppression of palmitate-induced Saa3 and Ccl2 gene expression by HDL is attenuated by glycation of the HDL by glycolaldehyde and malondialdehyde, but not glucose

After pre-incubation of adipocytes with CON-HDL, GAD-HDL or MDA-HDL for 6 hours, cells were challenged with PA for 24 hours. The PA-induced stimulation of Saa3 and Ccl2 gene expression was inhibited by pre-incubation with CON-HDL, GAD0.5-HDL, and MDA0.1-HDL. In contrast, pre-incubation with GAD1-HDL or MDA0.2-HDL did not prevent the increase in expression of Saa3 (Fig. 1A) and Ccl2 (Fig. 1B) relative to medium alone (non-stimulated control) and CON-HDL (p<0.0001), suggesting a concentration-dependent effect of in vitro modification. Glucose-modified HDL was partially able to prevent the increased Saa3 (Fig. 1C) and Ccl2 (Fig. 1D) expression elicited by PA, but did not exhibit a similar dose dependency as GAD- and MDA-modified HDL.

Figure 1: Suppression of palmitate-induced Saa3 and Ccl2 gene expression by HDL is attenuated by glycation of the HDL by glycolaldehyde and malondialdehyde, but not glucose.

Figure 1:

Open in a new tab

Adipocytes were pre-treated with 50 μg/mL of the various types of HDL for 6 hours, then with 250 μM of PA for 24 hours. RNA was extracted and converted into cDNA for analysis of expression of Saa3 (A and C) and Ccl2 (B and D). CON-HDL - control HDL; GAD - glycolaldehyde-modified HDL; MDA - malondialdehyde-modified HDL; GLU-HDL - glucose-modified HDL; – PA - palmitate. Data are expressed as mean ± SEM and are representative of at least 3 independent experiments. * p< 0.05 vs. PA alone, # p< 0.05 vs. CON-HDL + PA. ANOVA and Bonferroni post-hoc test.

3.2. HDL isolated from subjects with poorly controlled type I diabetes retains its anti-inflammatory properties on PA-stimulated adipocytes

To analyze the anti-inflammatory effects of HDL glycated in vivo, 3T3-L1 adipocytes were pre-exposed to three different concentrations of either ND-HDL or DM-HDL for 6 hours, followed by PA stimulation for 24 hours. The decrease in relative gene expression of Saa3 and Ccl2 was dose-dependent, as 50μg/mL of ND-HDL blunted the inflammatory response in these cells more effectively than 10 μg/mL (Fig. 2). Moreover, the same concentrations of DM-HDL inhibited the inflammatory response to the same degree as ND-HDL (p=0.16 for Saa3, p=0.10 for Ccl2). These results indicate that the anti-inflammatory properties of HDL from T1DM individuals in poor glycemic control are still effective relative to HDL from healthy non-diabetic controls.

Figure 2: HDL isolated from subjects with type 1 diabetes in poor glycemic control retains its anti-inflammatory properties on adipocytes stimulated with palmitate.

Figure 2:

Open in a new tab

DM-HDL or ND-HDL was isolated from the plasma of subjects with poorly T1DM (n=12) or normoglycemic healthy controls (n=9), respectively. Adipocytes were pre-treated with 10, 25, or 50 μg/mL of either ND-HDL or DM-HDL as described in the legend to Figure 1. Gene expression of Saa3 (A) and Ccl2 (B) was analyzed. Data, which are representative of at least 3 independent experiments, are expressed as mean ± SEM. DM-HDL: type 1 diabetes mellitus-HDL; ND-HDL: HDL from healthy non-diabetic subjects. Data. * p< 0.05 vs. 10 μg/mL. ANOVA and Bonferroni post-hoc test.

3.3. GAD- and MDA-HDL, but not DM-HDL, elicited reduced cholesterol efflux in adipocytes

Next, the potential of the various types of in vitro-modified HDL to mediate cholesterol efflux was assessed in 3T3-L1 adipocytes. Efflux mediated by GAD-HDL was impaired in a concentration-dependent manner. The differences in percentage of cholesterol efflux of GAD0.5-HDL or GAD1-HDL were 14.5%, and 19%, respectively, in comparison to CON-HDL (p<0.01; Fig. 3A). Similarly, cholesterol efflux mediated by MDA0.1-HDL and MDA0.2-HDL was impaired relative to CON-HDL, with a decrease in cholesterol efflux of 12.6% and 16.1% in comparison to control treatment (p<0.0006; Fig 3B). In contrast to HDL modified by either GAD or MDA, GLU-HDL retained its efflux capacities (p>0.05; Fig 3C). Moreover, DM-HDL did not exhibit impaired cholesterol efflux compared to ND-HDL. (p=0.69; Fig.3D)

Figure 3: Cholesterol efflux from adipocytes is reduced by GAD-HDL and MDA-HDL, but not GLU-HDL or DM-HDL.

Figure 3:

Open in a new tab

Cholesterol efflux capacity was tested in 3H-cholesterol-radiolabeled adipocytes incubated with or without 50 μg/mL of either GAD0.5 or GAD1-HDL (A), MDA0.1 or MDA0.2-HDL (B), GLU-HDL modified by exposure to either 5 (GLU5), 10 (GLU10) or 25 (GLU25) mmol glucose (C), and ND-HDL and DM-HDL (D). Data are expressed as mean ± SEM. * p< 0.05 vs. CON-HDL. ANOVA and Bonferroni post-hoc test in A-C. t test and Bonferroni post-hoc test in D.

3.4. Cholesterol efflux from macrophages was impaired after exposure to GAD-HDL and MDA-HDL, but not HDL modified by glucose or DM-HDL from subjects with poorly controlled type 1 diabetes

To understand whether these effects on efflux were adipocyte-specific, and considering the importance of macrophages in inflamed adipose tissue, we also tested ABCA1-dependent cholesterol efflux capacity of the various types of HDL in J774 macrophages pre-treated with cAMP to stimulate the expression of the cholesterol transporter, ABCA-1. HDL modified in vitro by GAD displayed lower cholesterol efflux capacity in a concentration-dependent manner compared to control HDL (−16.1% for GAD0.5 and −32.3% for GAD1.0; p<0.05; Fig. 4A). Cholesterol efflux mediated by MDA0.1-HDL was unchanged, while impairment was observed with MDA0.2-HDL (−15.6%; p<0.05; Fig.4B). In contrast, cholesterol efflux mediated by HDL modified by glucose was no different than control HDL (Fig. 4C). Moreover, the percentage of cholesterol efflux mediated by DM-HDL did not differ from ND-HDL (ND-HDL 5.9% ± 0.2 vs DM-HDL 6.2% ± 0.2), demonstrating that the cholesterol efflux capacity of DM-HDL also was not impaired in this assay (Fig. 4D).

Figure 4: Cholesterol efflux from macrophages is reduced by GAD-HDL and MDA-HDL, but not GLU-HDL or DM-HDL.

Figure 4:

Open in a new tab

Cholesterol efflux capacity was tested in 3H-cholesterol-radiolabeled macrophages incubated in media with or without 4 μg/mL of either GAD-HDL (A), MDA-HDL (B), GLU-HDL (C), and DM-HDL (D). Symbols are the same as in Figure 3. Data are expressed as mean ± SEM. t test and Bonferroni post-hoc test.

3.5. Lysine adducts in APOA1 and APOA2 in HDL glycated in vitro by GAD, MDA, or glucose

Mass spectrometric-based proteomic analysis was performed to evaluate post-translational modification adducts present in APOA1 and APOA2 from the HDL preparations glycated in vitro. Carboxymethyl-lysine (CML) was a major glycated lysine adduct in GAD-modified HDL; six lysine residues in APOA1 (K23, K96, K195, K206, K226, and K238) were modified by GAD to form CML. We observed a concentration dependent increase in the yield of CML at the three most modified lysine residues, K206, K23, and K195 (Fig. 5A). The major sites in APOA2 for GAD modification to form CML were K39, K44, and K55 and the product yields also were concentration dependent (Fig. 5B). In APOA1, among the five lysine residues (K59, K107, K140, K195 and K226) modified by MDA to form N-propenal-lysine adduct, lysines K140 and 107 were the most modified residues (Fig. 5C), while in APOA2, K55 is the major site for that MDA adduct (Fig. 5D). In MDA-modified HDL, MS analysis revealed the presence of three strong cross-linking products (Lys-MDA-Lys) within APOA1 (K12-K94, K77-K195, and K118-K140) (Fig. 5E) and one major cross-linking product in APOA2 (K23-K30) (Fig. 5F). The product yields of N-propenal-lysine and Lys-MDA-Lys were concentration dependent with increasing concentrations of MDA (Fig. 5C, 5D, 5E, and 5F). To determine whether glucose modified APOA1 and APOA2 in HDL, HDL was incubated with various concentrations of glucose for one week and MS was used to detect and quantify fructose-lysine (FL) adducts. LC-MS/MS analysis revealed that three major sites (K195, K206, and K226) in APOA1 and one major site (K55) in APOA2 were modified by glucose to form FL adducts in a glucose concentration-dependent manner (Figs. 5G and 5H). We also searched for arginine modification products potentially generated by methylglyoxal. However, no such products were identified in any of the in vitro modified HDL samples.

Figure 5: CML and MDA adducts are increased in distinct lysine residues of both APOA1 and APOA2 from GAD-HDL and MDA-HDL in a concentration dependent manner.

Figure 5:

Open in a new tab

HDL from normoglycemic healthy subjects was modified by either GAD, MDA or glucose at the indicated concentrations. Samples were digested and analyzed by targeted LC-MS/MS analysis. Changes in APOA1 (A, C, E and G) and APOA2 (B, D, F and H) lysine adducts of CML (A and B), MDA-lysine (C and D), Lys-MDA-Lys cross-links (E and F), and glucose-related adducts (fructose-lysine) (G and H) were observed.

3.6. Both APOA1 and APOA2 in HDL from subjects with type 1 diabetes in poor glycemic control contain fructose-lysine, resembling that of HDL modified in vitro with glucose but not with GAD or MDA

Targeted MS was used to compare the nature of APOA1 and APOA2 lysine adducts in HDL between patients with poorly controlled T1DM and healthy control subjects. Similar to our in vitro studies using glucose modified HDL, we identified FL adducts at four lysine residues in APOA1 (K12, K195, K206, and K226) and at one lysine residue in APOA2 (K55). Importantly, the levels of FL in APOA1 of HDL isolated from the patients with DM were two times higher than those from healthy control subjects (Fig. 6A, 6B, 6C, and 6D) and the levels of FL adduct in APOA2 of DM-HDL were 3 times higher than those of ND-HDL (Fig. 6E). In HDL isolated from the DM patients or from healthy control subjects, we also identified CML product at three lysine residues in APOA1 (K206, K208, and K226) but not in APOA2. However, levels of CML in APOA1 of HDL isolated from DM patients were significantly lower than that from control subjects (Fig. 6F, 6G, and 6H). We were unable to identify MDA adducts (N-propenal lysine or Lys-MDA-Lys) in great abundance in HDL. As with the in vitro modified samples, no arginine modification products formed by methylglyoxal were identified in any of the HDL samples from the control or T1DM subjects.

Figure 6. Post-translational changes in lysine residues observed in DM-HDL resemble those found in glucose-modified HDL.

Figure 6.

Open in a new tab

DM-HDL or ND-HDL were isolated from the plasma of poorly controlled T1DM (closed squares) or non-diabetic healthy control subjects (closed circles). Samples were digested and analyzed by targeted LC-MS/MS analysis. Major changes in lysine residues of APOA1 associated to glucose (A-E) and CML (F-H) in ND-HDL and DM-HDL were identified. Data are expressed as mean ± SEM. * p< 0.05 vs. ND-HDL. t test and Bonferroni post-hoc test.

4. Discussion

Chronic hyperglycemia is a hallmark of both T1DM and T2DM. Chronic hyperglycemia results in the generation of reactive aldehydes, which leads to the formation of AGEs and/or ALEs, which in turn have been shown to associate with diabetes-related complications (Brownlee, 2001; Nobecourt et al., 2010; Semba, Bandinelli, Sun, Guralnik, & Ferrucci, 2009; Semba, Fink, Sun, Windham, & Ferrucci, 2010; Semba et al., 2015). We previously have shown that HDL from lean mice and healthy human subjects inhibits the expression of chemokine and cytokine genes induced by PA in adipocytes (Han et al., 2016). Glycation of HDL in vitro or in vivo has been associated with its loss of function (Kashyap et al., 2018; Liu et al., 2012; Nobecourt et al., 2010). Therefore, we hypothesized that glycation of HDL might also impair the anti-inflammatory properties of HDL. We observed that pre-incubation of adipocytes with HDL modified in vitro by either GAD or MDA impaired this anti-inflammatory property of HDL in PA-stimulated adipocytes, an effect that was not observed with HDL modified with glucose or HDL isolated from T1DM subjects in poor glycemic control.

In vitro glycation of HDL using MDA or GAD exhibited an anti-inflammatory effect in adipocytes in a dose-dependent manner, with lower levels of glycation exhibiting comparable anti-inflammatory activity as control HDL. HDL modified in vitro by glucose inhibited the inflammatory response elicited by PA to a lesser extent to what we observed with control HDL. Although statistical differences between CON-HDL, GLU5-HDL and GLU25-HDL were observed in the expression of Saa3 (Fig. 1C), this effect did not appear to be glucose concentration-dependent. Others have found similar results using different cell types. Nobércourt et al. observed increased protein expression of adhesion molecules VCAM-1 and ICAM-1 when human coronary endothelial cells were pre-incubated with reconstituted HDL (rHDL) containing glycated HDL and challenged with TNF-α, in comparison to control rHDL (Nobecourt et al., 2010).

While our in vitro data showed that glycation-related modifications in HDL led to dysfunctional anti-inflammatory properties on adipocytes, these findings were not supported by the clinical component of our study, which was designed to determine whether the changes we observed in vitro could be emulated in vivo. HDL isolated from subjects with T1DM in poor glycemic control functioned normally with respect to its anti-inflammatory capacity in adipocytes. Importantly, the differences observed in the anti-inflammatory properties between HDL isolated from patients with T1DM and HDL glycated in vitro highlights the appreciation that changes made artificially in vitro systems may not necessarily reflect changes seen in vivo.

In the present study, we also evaluated the cholesterol efflux capacity of various glycated HDL preparations from adipocytes and showed that all types of HDL modified in vitro reduced cholesterol efflux compared to CON-HDL. We also found that cholesterol efflux in J774 macrophages was impaired after exposure to GAD-HDL and MDA-HDL, but not HDL modified by glucose. Moreover, HDL modified from poorly controlled T1DM subjects failed to replicate our findings with cholesterol efflux using in vitro modified HDL, again raising the question of whether similar modification occurs in vivo and in vitro. Others have reported similar findings for cholesterol efflux in other cell types. Our finding that GAD-HDL reduces cholesterol efflux capacity from J774 macrophages confirms the results from a previous study (Domingo-Espin, Nilsson, Bernfur, Del Giudice, & Lagerstedt, 2018). Hoang et al. suggested that worsening of cholesterol efflux capacity mediated by in vitro-modified HDL in HeLa and THP-1 cells is a result of alterations in APOA-I caused by advanced glycation (Hoang et al., 2007). A similar outcome was observed in RAW264.7 macrophages incubated with HDL from subjects with T2DM (Kashyap et al., 2018), the most significant effect of glycation of HDL being attributed to ABCA1-dependent cholesterol efflux. Consistent with our findings, cholesterol efflux capacity from THP-1 macrophages was unchanged in alloxan-induced DM mice compared to control mice despite glycation of HDL. The reduction of total macrophage-to-feces reverse cholesterol transport in this animal model was attributed to an impairment of HDL uptake in the liver mediated by the receptor SR-BI (de Boer et al., 2012). However, HDL from subjects with T1DM did show reduced efflux capacity irrespective of glycemic control (Manjunatha et al., 2016).

To evaluate whether in vitro glycation of HDL differs from what occurs in vivo, we used targeted proteomic methods to characterize and compare the lysine residues of APOA1 and APOA2. Our data show that lysine residues in APOA1 and APOA2 from HDL are modified by either GAD or MDA in a concentration-dependent manner. However, these changes are not mirrored by those found in HDL from our subjects with T1DM. When HDL was incubated with increasing concentrations of glucose in vitro, we observed increased fructose-lysine product at K195, K12, K226, and K206 in APOA1 and at K55 in APOA2. Interestingly, among these residues, only K23, K107, and K12 had been reported in the literature as modifiable sites by glycation in HDL and APOA1 incubated with glucose in vitro as well as in HDL from subjects with T2DM (Kashyap et al., 2018; Liu et al., 2018). Our subjects with T1DM in poor glycemic control had higher concentrations of fructose-lysine product at K195, K226, K206, and K12 in APOA1 and at K55 in APOA2 in HDL than in HDL from healthy control subjects, which mirrors what was found in the samples modified by glucose in vitro. The only shared residue in APOA1 in GAD and MDA-modified HDL was K195. The high reactivity of K195 could be explained by its location in an unstable C-terminal loop domain of APOA1, which is important for lipid binding. Indeed, K195 is a major site of modification and it forms cross-linking with several other lysine residues when lipid free APOA1 was exposed to MDA (Shao et al., 2010). However, replacement of K195 in APOA1 with cysteine did not affect the apolipoprotein’s ability to transport cholesterol, indicating that K195 itself does not play a particularly important role in terms of cholesterol efflux capacity of APOA1 (Cooke et al., 2018). Although previous studies have shown that modification of HDL in vitro with methylglyoxal altered the ability of the lipoprotein to facilitate cholesterol efflux (Domingo-Espin et al., 2018) and HDL modified by methylglyoxal was detected in subjects with T2DM (Godfrey, Yamada-Fowler, Smith, Thornalley, & Rabbani, 2014), we were unable to detect this form of arginine modification in HDL isolated from control or T1DM subjects in our study.

We and others previously have shown that the presence of SAA on HDL can affect its anti-inflammatory properties (Han et al., 2016; Tolle et al., 2012) and others have shown that SAA can affect the ability of the cholesterol efflux capacity of HDL (Vaisar et al., 2015). Therefore, we intentionally and specifically chose to study individuals with T1DM, thereby minimizing the systemic inflammation associated with obesity, which is associated with increased levels of SAA (Sjoholm et al., 2005; Tannock et al., 2005). The presence of higher concentrations of SAA in obese subject with T2DM might explain the difference between our finding in T1DM and those from a study that demonstrated that the anti-inflammatory properties of glycated APOA1 from patients with T2DM was impaired (Nobecourt et al., 2010). Evaluation of the effect of HDL from subjects with T2DM and/or cardiovascular disease addresses a very important but different question to that posed in the current study, and is beyond the scope of the present investigation.

A minor limitation of our study is that the control subjects were fasted overnight, whereas the T1DM subjects were not. Because of the requirement that people with T1DM eat shortly after insulin administration, it would have been very difficult to obtain fasted plasma from these subjects as outpatients. However, since the HDL from these subjects did not shown any functional abnormalities in our assay, and showed no protein differences from HDL modified in vitro by exposure to glucose, any potential addition from dietary AGEs and ALEs could not have made any major contribution.

Although we did not investigate other metabolic derangements such as chronic kidney disease, obesity and dyslipidemia that could also potentially lead to post-translational modifications of proteins in vivo, (Chatterjee & Thakur, 2018; Dunne, Overbergh, Purcell, & Mathieu, 2012), our study nonetheless suggests that in vitro modification of HDL by glycation may not accurately represent what happens to HDL in vivo. Our findings raise the question of whether the concentration of aldehydes used to modify these lipoproteins in vitro reflects changes that occur in vivo. In addition, the duration of exposure of HDL to glucose in vivo may result in different post-translational modifications of HDL than those achieved by in vitro exposure to GAD, MDA or glucose. However, the changes observed in vivo closely resembled those seen with modification by exposure to glucose in vitro. Another possible confounder is that most diabetic subjects in this study were being treated with statins, as is the standard of practice for most adults with diabetes. Statins may induce shedding of the receptor for advanced glycation end-products (RAGE) to generate its soluble form (sRAGE) (Quade-Lyssy, Kanarek, Baiersdorfer, Postina, & Kojro, 2013), the levels of which could increase in plasma. It has been suggested that increased plasma levels of sRAGE in T1DM might confer protection against cell damage together with elimination of circulating MDA, thereby decreasing the likelihood of post-translational changes of protein elicited by glycation (Reis et al., 2012). This may in part explain why HDL from the subjects with diabetes in the present study maintained its anti-inflammatory and cholesterol efflux properties. In conclusion, our findings suggest that the changes observed in glycation of HDL-associated APOA1 in poorly controlled patients with T1DM are insufficient to result in the levels of dysfunction of HDL seen with AGE or ALE modifications in vitro, and points out the potential downside of extrapolating data observed in vitro with what actually occurs physiologically.

Highlights.

  • In vitro glycation of HDL affects its ability to inhibit inflammation in adipocytes

  • HDL glycated in vitro also has impaired cholesterol efflux capacity

  • However HDL from poorly controlled type 1 patients functions normally

  • In vitro and in vivo glycation result in different modifications detected by MS

  • Glycation of HDL in vitro differs from that observed in vivo

5. Acknowledgments

This study was supported by NIH grant HL 092969. The authors thank Carl Storey and Angela Irwin from the UW Diabetes Institute for assistance with the cholesterol efflux assays.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

6. References:

  1. Ahmad S, Siddiqui Z, Rehman S, Khan MY, Khan H, Khanum S, … Saeed M (2017). A Glycation Angle to Look into the Diabetic Vasculopathy: Cause and Cure. Curr Vasc Pharmacol, 15(4), 352–364. doi: 10.2174/1570161115666170327162639 [DOI] [PubMed] [Google Scholar]
  2. Brownlee M (2001). Biochemistry and molecular cell biology of diabetic complications. Nature, 414(6865), 813–820. [DOI] [PubMed] [Google Scholar]
  3. Chatterjee B, & Thakur SS (2018). Investigation of post-translational modifications in type 2 diabetes. Clin Proteomics, 15, 32. doi: 10.1186/s12014-018-9208-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cooke AL, Morris J, Melchior JT, Street SE, Jerome WG, Huang R, … Davidson WS (2018). A thumbwheel mechanism for APOA1 activation of LCAT activity in HDL. J Lipid Res, 59(7), 1244–1255. doi: 10.1194/jlr.M085332 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Curtiss LK, & Witztum JL (1985). Plasma apolipoprotein A-I, A-II, B, C-I and E are glycosylated in hyperglycemic diabetic subjects. Diabetes, 34, 452–461. [DOI] [PubMed] [Google Scholar]
  6. de Boer JF, Annema W, Schreurs M, van der Veen JN, van der Giet M, Nijstad N, … Tietge UJ (2012). Type I diabetes mellitus decreases in vivo macrophage-to-feces reverse cholesterol transport despite increased biliary sterol secretion in mice. J Lipid Res, 53(3), 348–357. doi: 10.1194/jlr.M018671 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. de la Llera-Moya M, Drazul-Schrader D, Asztalos BF, Cuchel M, Rader DJ, & Rothblat GH (2010). The ability to promote efflux via ABCA1 determines the capacity of serum specimens with similar high-density lipoprotein cholesterol to remove cholesterol from macrophages. Arterioscler Thromb Vasc Biol, 30(4), 796–801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Domingo-Espin J, Nilsson O, Bernfur K, Del Giudice R, & Lagerstedt JO (2018). Site-specific glycations of apolipoprotein A-I lead to differentiated functional effects on lipid-binding and on glucose metabolism. Biochim Biophys Acta Mol Basis Dis, 1864(9 Pt B), 2822–2834. doi: 10.1016/j.bbadis.2018.05.014 [DOI] [PubMed] [Google Scholar]
  9. Dunne JL, Overbergh L, Purcell AW, & Mathieu C (2012). Posttranslational modifications of proteins in type 1 diabetes: the next step in finding the cure? Diabetes, 61(8), 1907–1914. doi: 10.2337/db11-1675 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Godfrey L, Yamada-Fowler N, Smith J, Thornalley PJ, & Rabbani N (2014). Arginine-directed glycation and decreased HDL plasma concentration and functionality. Nutr Diabetes, 4, e134. doi: 10.1038/nutd.2014.31 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gomes DJ, Velosa AP, Okuda LS, Fusco FB, da Silva KS, Pinto PR, … Passarelli M (2016). Glycated albumin induces lipid infiltration in mice aorta independently of DM and RAS local modulation by inducing lipid peroxidation and inflammation. J Diabetes Complications. doi: 10.1016/j.jdiacomp.2016.07.001 [DOI] [PubMed] [Google Scholar]
  12. Griffiths K, Pazderska A, Ahmed M, McGowan A, Maxwell AP, McEneny J, … McKay GJ (2017). Type 2 Diabetes in Young Females Results in Increased Serum Amyloid A and Changes to Features of High Density Lipoproteins in Both HDL2 and HDL3. J Diabetes Res, 2017, 1314864. doi: 10.1155/2017/1314864 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Han CY, Tang C, Guevara ME, Wei H, Wietecha T, Shao B, … Chait A (2016). Serum amyloid A impairs the antiinflammatory properties of HDL. J Clin Invest, 126(1), 266–281. doi: 10.1172/JCI83475 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hoang A, Murphy AJ, Coughlan MT, Thomas MC, Forbes JM, O’Brien R, … Sviridov D (2007). Advanced glycation of apolipoprotein A-I impairs its anti-atherogenic properties. Diabetologia, 50(8), 1770–1779. doi: 10.1007/s00125-007-0718-9 [DOI] [PubMed] [Google Scholar]
  15. Kamtchueng Simo O, Ikhlef S, Berrougui H, & Khalil A (2017). Advanced glycation end products affect cholesterol homeostasis by impairing ABCA1 expression on macrophages. Can J Physiol Pharmacol, 95(8), 977–984. doi: 10.1139/cjpp-2017-0170 [DOI] [PubMed] [Google Scholar]
  16. Kanter JE, Shao B, Kramer F, Barnhart S, Shimizu-Albergine M, Vaisar T, … Bornfeldt KE (2019). Increased apolipoprotein C3 drives cardiovascular risk in type 1 diabetes. J Clin Invest, 130, 4165–4179. doi: 10.1172/JCI127308 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kashyap SR, Osme A, Ilchenko S, Golizeh M, Lee K, Wang S, … Kasumov T (2018). Glycation Reduces the Stability of ApoAI and Increases HDL Dysfunction in Diet-Controlled Type 2 Diabetes. J Clin Endocrinol Metab, 103(2), 388–396. doi: 10.1210/jc.2017-01551 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kumon Y, Suehiro T, Itahara T, Ikeda Y, & Hashimoto K (1994). Serum amyloid A protein in patients with non-insulin-dependent diabetes mellitus. Clin Biochem, 27, 469–473. [DOI] [PubMed] [Google Scholar]
  19. Liu D, Ji L, Zhang D, Tong X, Pan B, Liu P, … Zheng L (2012). Nonenzymatic glycation of high-density lipoprotein impairs its anti-inflammatory effects in innate immunity. Diabetes Metab Res Rev, 28(2), 186–195. doi: 10.1002/dmrr.1297 [DOI] [PubMed] [Google Scholar]
  20. Liu D, Ji L, Zhao M, Wang Y, Guo Y, Li L, … Zheng L (2018). Lysine glycation of apolipoprotein A-I impairs its anti-inflammatory function in type 2 diabetes mellitus. J Mol Cell Cardiol, 122, 47–57. doi: 10.1016/j.yjmcc.2018.08.001 [DOI] [PubMed] [Google Scholar]
  21. Manjunatha S, Distelmaier K, Dasari S, Carter RE, Kudva YC, & Nair KS (2016). Functional and proteomic alterations of plasma high density lipoproteins in type 1 diabetes mellitus. Metabolism: Clinical and Experimental, 65(9), 1421–1431. doi: 10.1016/j.metabol.2016.06.008 [DOI] [PubMed] [Google Scholar]
  22. Newsholme P, Cruzat VF, Keane KN, Carlessi R, & de Bittencourt PI Jr. (2016). Molecular mechanisms of ROS production and oxidative stress in diabetes. Biochem J, 473(24), 4527–4550. doi: 10.1042/BCJ20160503C [DOI] [PubMed] [Google Scholar]
  23. Nobecourt E, Tabet F, Lambert G, Puranik R, Bao S, Yan L, … Rye KA (2010). Nonenzymatic glycation impairs the antiinflammatory properties of apolipoprotein A-I. Arterioscler Thromb Vasc Biol, 30(4), 766–772. doi: 10.1161/ATVBAHA.109.201715 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Ott C, Jacobs K, Haucke E, Navarrete Santos A, Grune T, & Simm A (2014). Role of advanced glycation end products in cellular signaling. Redox Biol, 2, 411–429. doi: 10.1016/j.redox.2013.12.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Pu LJ, Lu L, Zhang RY, Du R, Shen Y, Zhang Q, … Shen WF (2013). Glycation of apoprotein A-I is associated with coronary artery plaque progression in type 2 diabetic patients. Diabetes Care, 36(5), 1312–1320. doi: 10.2337/dc12-1411 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Quade-Lyssy P, Kanarek AM, Baiersdorfer M, Postina R, & Kojro E (2013). Statins stimulate the production of a soluble form of the receptor for advanced glycation end products. J Lipid Res, 54(11), 3052–3061. doi: 10.1194/jlr.M038968 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Reddy P, Lent-Schochet D, Ramakrishnan N, McLaughlin M, & Jialal I (2019). Metabolic syndrome is an inflammatory disorder: A conspiracy between adipose tissue and phagocytes. Clin Chim Acta, 496, 35–44. doi: 10.1016/j.cca.2019.06.019 [DOI] [PubMed] [Google Scholar]
  28. Reis JS, Veloso CA, Volpe CM, Fernandes JS, Borges EA, Isoni CA, … Nogueira-Machado JA (2012). Soluble RAGE and malondialdehyde in type 1 diabetes patients without chronic complications during the course of the disease. Diab Vasc Dis Res, 9(4), 309–314. doi: 10.1177/1479164111436316 [DOI] [PubMed] [Google Scholar]
  29. Semba RD, Bandinelli S, Sun K, Guralnik JM, & Ferrucci L (2009). Plasma carboxymethyl-lysine, an advanced glycation end product, and all-cause and cardiovascular disease mortality in older community-dwelling adults. J Am Geriatr Soc, 57(10), 1874–1880. doi: 10.1111/j.1532-5415.2009.02438.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Semba RD, Fink JC, Sun K, Windham BG, & Ferrucci L (2010). Serum carboxymethyl-lysine, a dominant advanced glycation end product, is associated with chronic kidney disease: the Baltimore longitudinal study of aging. J Ren Nutr, 20(2), 74–81. doi: 10.1053/j.jrn.2009.08.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Semba RD, Sun K, Schwartz AV, Varadhan R, Harris TB, Satterfield S, … Health ABCS (2015). Serum carboxymethyl-lysine, an advanced glycation end product, is associated with arterial stiffness in older adults. J Hypertens, 33(4), 797–803; discussion 803. doi: 10.1097/HJH.0000000000000460 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Shao B, & Heinecke JW (2008). Using tandem mass spectrometry to quantify site-specific chlorination and nitration of proteins: model system studies with high-density lipoprotein oxidized by myeloperoxidase. Methods Enzymol, 440, 33–63. doi: 10.1016/S0076-6879(07)00803-8 [DOI] [PubMed] [Google Scholar]
  33. Shao B, Pennathur S, Pagani I, Oda MN, Witztum JL, Oram JF, & Heinecke JW Modifying apolipoprotein A-I by malondialdehyde, but not by an array of other reactive carbonyls, blocks cholesterol efflux by the ABCA1 pathway. J Biol Chem, 285(24), 18473–18484. doi: M110.118182 [pii] 10.1074/jbc.M110.118182 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Shao B, Pennathur S, Pagani I, Oda MN, Witztum JL, Oram JF, & Heinecke JW (2010). Modifying apolipoprotein A-I by malondialdehyde, but not by an array of other reactive carbonyls, blocks cholesterol efflux by the ABCA1 pathway. J Biol Chem, 285(24), 18473–18484. doi: 10.1074/jbc.M110.118182 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Sjoholm K, Palming J, Olofsson LE, Gummesson A, Svensson PA, Lystig TC, … Carlsson LM (2005). A microarray search for genes predominantly expressed in human omental adipocytes: adipose tissue as a major production site of serum amyloid A. J Clin Endocrinol Metab, 90(4), 2233–2239. [DOI] [PubMed] [Google Scholar]
  36. Tannock LR, O’Brien KD, Knopp RH, Retzlaff B, Fish B, Wener MH, … Chait A (2005). Cholesterol feeding increases C-reactive protein and serum amyloid A levels in lean insulin-sensitive subjects. Circulation, 111(23), 3058–3062. [DOI] [PubMed] [Google Scholar]
  37. Tolle M, Huang T, Schuchardt M, Jankowski V, Prufer N, Jankowski J, … van der Giet M (2012). High-density lipoprotein loses its anti-inflammatory capacity by accumulation of pro-inflammatory-serum amyloid A. Cardiovasc Res, 94(1), 154–162. doi: 10.1093/cvr/cvs089 [DOI] [PubMed] [Google Scholar]
  38. Vaisar T, Tang C, Babenko I, Hutchins P, Wimberger J, Suffredini AF, & Heinecke JW (2015). Inflammatory Remodeling of the HDL Proteome Impairs Cholesterol Efflux Capacity. J Lipid Res. doi: 10.1194/jlr.M059089 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Wong NKP, Nicholls SJ, Tan JTM, & Bursill CA (2018). The Role of High-Density Lipoproteins in Diabetes and Its Vascular Complications. Int J Mol Sci, 19(6). doi: 10.3390/ijms19061680 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES