A Novel ELISA System for Measuring Modified LDL-Adiponectin Complex

Abstract

Aim: Adiponectin is an anti-diabetic and anti-atherogenic protein secreted primarily from adipose tissue. Adiponectin and modified LDL (mLDL) form a complex to modulate their biological activity. To elucidate the significance of the complex formation, we analyzed its effects on vascular tissue and developed and verified novel quantifying methods for adiponectin.

Methods: To study the significance of the mLDL-adiponectin complex (MAC) formation, we used the wire-myography method on rat mesenteric artery. We developed a method to measure MAC by using LOX-1 as the capture protein and anti-adiponectin antibody for detection. We compared serum MAC levels between hemodialysis patients and control subjects.

Results: Administering mLDL alone to rat mesenteric artery impaired endothelium-dependent vasorelaxation, whereas simultaneously administering adiponectin with mLDL protected rat mesenteric artery from the mLDL-induced impairment of vasorelaxation. This finding indicates MAC formation prevents endothelium from mLDL-induced dysfunction in tissue. Using our novel ELISA for MAC, we found that MAC was increasingly detectable depending on the doses of mLDL and adiponectinin vitro. In serum, hemodialysis patients showed a significantly higher ratio of MAC-high patients (higher than the median level of MAC) than did healthy controls. Furthermore, the MAC-high hemodialysis group had lower mLDL activity measured as LOX-1 ligand containing apoB.

Conclusion: Using our ELISA, we detected MAC in human serum that protected blood vessels from the deleterious effects of oxidized LDL.

Introduction

Adiponectin is an adipokine secreted mainly from adipose tissue ^{1 , 2)} but can be produced in skeletal muscle ³⁾ and cardiomyocytes ⁴⁾ . Encoded by the ADIPOQ (human 3q27) gene ⁵⁾ , adiponectin comprises 244 amino acids and has a molecular weight of 28 kDa. The molecule has a variable region, a collagen-like domain, and a globular domain that is structurally homologous to C1q ⁶⁾ . In the circulation, adiponectin exists as trimers, hexamers, and high-molecular-weight (HMW) multimers ^{7 , 8)} .

Adiponectin has anti-inflammatory and anti-atherogenic effects in addition to its effects on energy metabolism, which increase insulin sensitivity. Adiponectin may reduce C-reactive protein expression, inhibit NF-Kβ signaling and TNF-α secretion from macrophages ^{9 - 11)} , and suppress the activity of several atherogenic factors. Some studies suggest adiponectin inhibits the proliferation and migration of human aortic smooth muscle cells ^{9 , 12 , 13)} .

The adiponectin receptors, AdipoR1 and AdipoR2, were identified by Yamauchi et al. ¹⁴⁾ , and T-cadherin, another receptor for adiponectin, was identified by Hug et al. ¹⁵⁾ . In a mouse model, adiponectin localized in the arterial endothelium by binding to T-cadherin. In contrast, in T-cadherin-deficient mice, adiponectin accumulation in the arterial endothelium decreased, and adiponectin levels in the blood increased ¹⁶⁾ . Furthermore, in atherosclerotic lesions in ApoE-deficient mice, adiponectin accumulated both in the endothelium and in proliferating smooth muscle cells, colocalizing with T-cadherin; this adiponectin accumulation was not seen in T-cadherin/ApoE double-deficient mice, where more severe atherosclerotic lesions were noted ¹⁷⁾ . These findings suggest that the localization of adiponectin in vascular tissue via T-cadherin is responsible, at least in part, for the anti-arteriosclerosis effect of adiponectin.

Patients with coronary artery disease have low blood levels of adiponectin ^{18 , 19)} , and in patients with heart failure, the blood levels of adiponectin and atherogenic oxidized LDL (oxLDL) are inversely correlated ²⁰⁾ . Moreover, the levels of total and HMW adiponectin are inversely correlated with coronary artery disease, ischemic stroke, and peripheral artery disease ^{21 - 23)} . However, in several reports ^{24 - 26)} , the prognosis for cardiovascular disease was poor when adiponectin concentrations were high, indicating there may not be a simple linear correlation.

High concentrations of adiponectin in patients with chronic kidney disease are often accompanied by high concentrations of cystatin C, which can bind to adiponectin ²⁷⁾ . This binding reduces the clearance rate of adiponectin, thus increasing blood levels, and suppresses the physiological activity of adiponectin. These findings suggest that a substance that interacts with adiponectin may affect both its circulating levels and biological activity. We have previously shown that adiponectin interacts with and suppresses the activity of oxidatively modified LDL. Because modified LDL appears to play a pivotal role in the pathogenesis of arteriosclerosis, this interaction may attenuate the progression of atherosclerosis ²⁸⁾ .

Aim

We analyzed the significance of modified LDL-adiponectin complex (MAC) formation on the ex vivo effects of oxLDL in vascular tissues. Also, we have developed the first method for quantifying MAC and the physiological activity of adiponectin. Using human blood samples, we have examined the clinical significance of MAC formation.

Methods

Ex Vivo Vascular Function Experiments by Wire Myography

Four- to six-month-old virgin female Sprague Dawley rats were purchased from Charles River, Canada (Saint-Constant, QC, Canada) and were housed under a standard day and night cycle (10:14 h) with ad libitum access to food and water. All rats were acclimatized at least one week after arrival and then euthanized by exsanguination under inhaled isoflurane anesthesia. Second-order mesenteric arteries were isolated, divided into 2-mm pieces, and mounted on a single 40-µm wire. The segments were then incubated with or without oxLDL (50 µg/mL, Kalen Biomedical, LLC, Germantown, MD, USA), adiponectin (30 µg/mL, BioVendor), the anti-LOX-1 neutralizing antibody TS20 (10 µg/mL) ^{29 - 33)} , or a mouse immunoglobulin G isotype (IgG, 10 µg/mL, MilliporeSigma, Burlington, MA, USA) overnight (~16 h) at 37℃ in culture media (Medium 199 [M199, Gibco, Thermo Fisher] supplemented with 8% fetal bovine serum [MilliporeSigma], 1% penicillin-streptomycin [Gibco], and 0.1% gentamicin [MilliporeSigma]).

To assess vascular function, we mounted the mesenteric artery segments onto a wire myograph system (DMT, Copenhagen, Denmark) by using 40 µm wires in HEPES-buffered physiological saline solution (PSS; in mmol/L: 142 NaCl, 4.7 KCl, 1.17 MgSO₄, 4.7 CaCl₂, 1.18 K₂PO₄, 10 HEPES, and 5.5 glucose; pH 7.4), and the segments were then normalized (IC100 = 0.8; 13.3 kPa). Vessels were exposed to a 10 µmol/L dose of phenylephrine (MilliporeSigma) followed by a dose (3 µmol/L) of methylcholine (MCh; MilliporeSigma), which is an endothelium-dependent vasodilator, to test for endothelial integrity and function. The arteries were pre-constricted with U46619 (thromboxane A2 analogue, MilliporeSigma) at a previously determined EC₈₀ dose (0.16 µmol/L; data not shown), and endothelium-dependent vasodilation responses were assessed with a cumulative concentration-response curve to MCh (1×10⁻⁹ to 1×10⁻⁴ mol/L MCh). Finally, each artery was incubated with a high potassium salt solution (KPSS; 123 mmol/L) to assess non-receptor-mediated vasoconstriction responses. Data were analyzed using LabChart software (ADInstruments; Colorado Springs, CO, USA), and the maximum effect (Emax, computed as the point of maximal vasodilation for each n) was calculated.

Preparation of Recombinant T-cadherin Protein

Human T-cadherin cDNA (accession no. NM_001257.5) was amplified from human heart cDNA (Human MTC Panel I; Clontech Laboratories, Mountain View, CA, USA). The T-cadherin PCR product was digested with HindIII and NotI and cloned into the pSeq-Tag2(B) vector (Thermo Fisher Scientific, Waltham, MA, USA, V90020), with Myc and His tag fused at the C-terminus. An R138A mutation was inserted by inverse PCR using the KOD-Plus-Mutagenesis Kit (TOYOBO, Osaka, Japan, SMK-101). The DNA sequence was determined using the Big-Dye terminator v3.1 cycle sequencing kit (Thermo Fisher Scientific, 4337455). The expression vector was purified using EndoFree Plasmid Maxi Kit (Qiagen, Hilden, Germany, 12362) for transfection experiments.

The proteins were expressed using Expi293 Expression System (Thermo Fisher Scientific). In brief, Expi293 cells were cultured in a shake flask (37℃, 8% CO₂) set at 135 rpm. Each expression vector was transfected according to the procedures recommended by the manufacturer. Recombinant T-cadherin proteins were expressed for 3 days in the flask under the above culture conditions before harvesting. The recombinant proteins were purified from the culture medium using Ni Sepharose excel (Cytiva, Tokyo, Japan) with Poly-Prep Chromatography Columns (Bio-Rad Laboratories, Hercules, CA, US). Purified proteins were dialyzed against phosphate buffered saline (PBS) and sterilized by using a 0.2 µm filter. The quantity was determined by using the BCA assay (Thermo Fisher Scientific, 23228).

Construction of a novel ELISA for Adiponectin that Retains the Capability to Bind to T-Cadherin

Recombinant human T-cadherin (0.15 µg/well) dissolved in PBS (30 µL/well) was immobilized on 384-well plates by incubating overnight at 4℃. After the wells were washed twice with PBS, 1% (w/v) casein-Na blocking solution was added, and the plates were incubated for 2 h at 25℃. The wells were washed three times with PBS, and the plates were incubated for 2 h at 25℃ with 20 µL of the recombinant adiponectin (BioVendor R&D, Brno, Czech Republic, RD172023100) or serum diluted four times with HEPES-NaCl buffer (10 mM HEPES, 150 mM NaCl, pH 7.4) containing 1 mM CaCl₂ and 0.1 mM MgCl₂. The wells were washed three more times with PBS, and the plates were incubated for 1 h at 25℃ with 1.0 µg/mL anti-adiponectin antibody (BioVendor, RD181023100) in 1% (w/v) casein-Na blocking solution. After the wells were washed three times with PBS, the plates were incubated for 1 h at 25℃ with peroxidase-conjugated goat anti-rabbit IgG (Thermo Fisher Scientific, 31458) that was diluted 4000 times with 1% (w/v) casein-Na blocking solution. Then, the substrate solution (tetramethylbenzidine [TMB] solution, Bio-Rad Laboratories) was added to the plates after the wells were washed five times, and the plates were incubated at room temperature. The reaction was terminated with 2 M sulfuric acid. Peroxidase activity was determined by measuring absorbance at 450 nm.

Measurement of Modified LDL-Adiponectin Complex (MAC)

Recombinant human LOX-1 (0.15 µg/well) dissolved in PBS (30 µL/well) was immobilized on 384-well plates by incubating overnight at 4℃. After the wells were washed twice with PBS, 1% (w/v) casein-Na blocking solution was added, and the plates were incubated for 2 h at 25℃. After three washes with PBS, the plates were incubated for 2 h at room temperature with mixture of oxLDL and adiponectin. Detection was achieved using anti-adiponectin antibody in the same manner as described above.

For human serum measurements, serum was diluted 4-fold with HEPES-NaCl buffer containing 1 mM CaCl₂ and 0.1 mM MgCl₂, added to the plate after blocking, and incubated for 2 h.

Reactions for drawing standard curve were prepared by mixing 1 µg/mL of oxidized LDL and various concentrations of adiponectin. The optical density of the reaction for each sample was translated to the unit of µg adiponectin/mL by referring to the standard curve. Thus, the MAC concentration is expressed by adiponectin concentration bound to the fixed concentration of oxidized LDL with the unit of µg adiponectin/mL.

Measurement of LOX-1 Ligands Containing apoB (LAB)

Measurement of LOX-1 ligands containing apoB (LAB) in human serum samples was conducted as previously reported ³⁴⁾ . Briefly, recombinant human LOX-1 dissolved in PBS was immobilized on 384-well plates by incubating overnight at 4℃. After 3 washes with PBS, 3% BSA in HEPES-NaCl buffer was added, and the plates were incubated for 2 h at 25℃. After 3 washes with PBS, the plates were incubated for 2 h at room temperature with standard oxidized LDL or samples. Samples were prepared by 20-fold dilution of serum with EDTA-BSA-HEPES buffer (2 mmol/L EDTA, 5% BSA/HEPES-NaCl buffer), and standards were prepared by dilution of oxidized LDL with EDTA-BSA-HEPES buffer. After 3 washes with PBS, the plates were incubated for 1 h at room temperature with chicken monoclonal anti-ApoB antibody (HUC20) in EDTA-BSA-HEPES buffer. After 3 washes with PBS, the plates were incubated for 1 h at room temperature with peroxidase-conjugated donkey anti-chicken IgY (Merck, Darmstadt, Germany, AP194P) diluted 6000 times with EDTA-BSA-HEPES buffer. After 5 washes with PBS, TMB solution was added to the plates, which were incubated for 30 min at room temperature. The reaction was terminated with 2 M sulfuric acid. Peroxidase activity was determined by measuring absorbance at 450 nm.

Measurement of Oxidized LDL by Conventional ELISA

The serum levels of oxLDL were determined by using a conventional oxidized LDL ELISA kit (Mercodia, Uppsala, Sweden, 10-1143-01), which uses the monoclonal antibody 4E6 that specifically recognizes the three-dimensional epitope of the oxidized ApoB-100 molecule. The measurement was performed according to the manufacturer’s instructions. Briefly, 200 µL of assay buffer and 25 µL of 1/6561 diluted serum samples and standards were added to a 96-well plate, and the plates were incubated for 2 h at room temperature on a plate shaker. The plates were washed 6 times with wash buffer, 200 µL of enzyme conjugate solution was added into each well, and the plates were incubated for 1 h at room temperature with shaking on a plate shaker. The plates were washed again 6 times with wash buffer, and 200 µL of the substrate TMB was added. After the plates were incubated for 15 min at room temperature, 50 µL of stop solution was added into each well. The O.D. absorbance was read at 450 nm in a microplate reader. All solutions used in the experiments were from the kit.

Measurement of Total- and HMW-Adiponectin

The serum levels of total- and HMW-adiponectin were measured using ELISA kits (R&D Minneapolis, MN, USA, DRP300 for total-adiponectin, DHWAD0 for HMW-adiponectin) according to the manufacture’s protocol. Briefly, 50 µL of 100-fold diluted serum samples and standards were added to 96-well plates preloaded with 100 µL of assay diluent. The plates were incubated at 25℃ for 2h (DRP300) or 3h (DHWAD0). The well contents were discarded, and the plates were washed 4 times with 300 µL wash buffer. Then, 200 µL of prepared antibody was added to each well, and the plates were incubated at 25℃ for 2 h (DRP300) or 1 h (DHWAD0) and washed 4 times with wash buffer. The substrate solution (200 µL) was added to each well, and the plates were incubated at room temperature in the dark for 30 min. Then, 50 µL of stop solution was added to each well, and the absorbance was read at 450 nm on a microplate reader. All reagents were provided in the ELISA kits.

Subjects

For the analysis of correlations between total-adiponectin, HMW-adiponectin, and Tcad-adiponectin, serum samples were obtained from 40 patients who received outpatient treatment at Shinshu University Hospital between January 2023 and April 2023.

For the analysis of MAC and its clinical significance, serum samples were obtained from a total of 52 patients with chronic kidney disease who underwent hemodialysis at Shinshu University Hospital between January 2023 and April 2023. We excluded 17 patients for a history of malignancy. Stocked serum samples from age- and sex-matched subjects (n = 35) with no history of dyslipidemia, diabetes mellitus, or malignancy among those who underwent medical checkups at Yodakubo Hospital were used as healthy controls. The subjects were divided into MAC-high and -low groups according to the median level of MAC.

Serum samples were stored in aliquots at −80℃ until use. The racial background of the subjects was uniformly Japanese. The studies were conducted according to the guidelines of the Declaration of Helsinki and were approved by the Human Ethics Committees of Shinshu University School of Medicine (Matsumoto, Japan) (approval no. 5731).

Statistical Analysis

For rat vascular experiments, continuous variable data are shown as the mean±standard error of the mean (SEM). The significance of the differences was analyzed by two-way ANOVA with Tukey post-hoc test or Sidak post hoc test. For blood sample measurements, data are shown as the mean±standard deviation (SD). The correlations between two continuous variables were analyzed by correlation analysis. The significance of the differences between the hemodialysis patients and control groups was analyzed using the unpaired Student’s t-test or the chi-square test.

All statistical analyses were performed using Graph Pad Prism 9.0 (GraphPad Software, Inc., La Jolla, CA) or StatFlex ver. 7.0.11 (Artech Co., Ltd., Osaka, Japan) software. Statistical significance was set at p＜0.05.

Results

Adiponectin Interferes With oxLDL-induced Impairment of Vasorelaxation Reaction

We used a rat mesenteric artery preparation to examine whether MAC formation suppresses the oxLDL effects on vascular function. As we have previously shown ³⁵⁾ , treatment of the mesenteric artery preparation with oxLDL impaired the MCh-induced vasorelaxation response. Here, we found that simultaneously applying adiponectin with oxLDL significantly blocked the oxLDL effect on vasorelaxation ( Fig.1A, B ) . Adiponectin alone did not affect the vasorelaxation response. Furthermore, application of anti-LOX-1 neutralizing antibody TS20 ^{29 - 33)} in addition to adiponectin did not show an additional effect on the oxLDL-induced vasorelaxation impairment ( Fig.1C, D ) , confirming that the effect of adiponectin occurs by blocking the binding of oxLDL to LOX-1.

Fig.1. Effect of adiponectin (AN) on endothelium-dependent vasodilation in rat mesenteric arteries treated with oxLDL.

(A) Cumulative concentration response curve of methylcholine-induced vasodilation in pre-constricted (with U46619) mesenteric arteries incubated overnight without stimulus (white), 50 µg/mL of oxLDL (red), 30 µg/mL of AN (green), or 30 µg/mL of AN plus 50 µg/mL of oxLDL (blue). (B) Summary graph of the maximum response (Emax) obtained from concentrations curves shown in (A). Vessels were obtained from rats (n = 6-13). (C) Cumulative concentration response curve of methylcholine-induced vasodilation in pre-constricted (with U46619) mesenteric arteries incubated overnight without stimulus (white) or with 50 µg/mL of oxLDL (red), 30 µg/mL of AN plus 50 µg/mL of oxLDL (blue), 30 µg/mL of AN plus 10 µg/mL of TS20 (purple), or 10 µg/mL of the IgG isotype (black). (D) Summary graph of the maximum response (Emax) obtained from concentrations curves shown in C. Vessels from n = 6-7 rats. Values are shown as the mean±SEM. ^＊p＜0.05; ^＊＊p＜0.01; ^＊＊＊p＜0.001. ns, non-significant.

Development of an ELISA System for Detecting T-cadherin Bindable Adiponectin (Tcad-AN) and Modified LDL-adiponectin Complex (MAC)

We have reported that the physiological activity of adiponectin may be affected by the presence of atherogenic LDL ²⁸⁾ . Thus, we developed a novel ELISA system to measure the binding activity of circulating adiponectin to T-cadherin. Because a mouse study showed that the precursor form, prodomain containing T-cadherin, may show an enhanced binding of adiponectin ³⁶⁾ , we first designed cDNA for prodomain-containing human T-cadherin for expression in mammalian cells. To avoid the removal of the prodomain, we introduced a mutation to the T-cadherin cDNA that converts arginine at position 138, which should be recognized by the processing enzyme, to alanine (R138A) ( Fig.2A ) . We also designed the cDNA for a prodomain deletion mutant (Δprodomain) to obtain pure T-cadherin without the prodomain since wild type T-cadherin cDNA produces the mixture of prodomain-containing and prodomain-absent T-cadherin in our mammalian cell-expression system ( Fig.2A ) . SDS-PAGE and western blot analysis of the purified T-cadherin under non-reducing conditions showed that cDNA for R138A-T-cadherin produced only prodomain-containing T-cadherin and that for Δprodomain produced only mature prodomain-absent T-cadherin, respectively ( Fig.2B ) . The bands around 200-250kDa indicate dimers of T-cadherin interacted via the EC1-EC2 domains as previously reported ³⁷⁾ .

Fig.2. Construction of a novel ELISA for adiponectin (AN) utilizing T-cadherin as a capture protein.

(A) Schematic representation of recombinant human T-cadherin proteins fused with Myc-His-tag. (B) Coomassie brilliant blue (CBB) staining and western blot analysis of purified T-cadherin proteins. (C) Binding of recombinant AN to immobilized T-cadherin in ELISA was determined by anti-AN antibody. (D) Binding of recombinant AN−oxLDL complex to immobilized LOX-1 in ELISA was determined by anti-AN antibody.

s.p., signal peptide; EC, extracellular cadherin.

Next, we constructed a sandwich ELISA system using the prepared T-cadherin proteins and an anti-adiponectin polyclonal antibody ( Fig.2C ) . Recombinant T-cadherin was used as the capture protein. We added adiponectin (serially diluted, 0.156 – 5 µg/mL) to the wells. Anti-human adiponectin polyclonal antibody was used for detection. Adiponectin bound to both prodomain-containing and prodomain-absent T-cadherin in a dose-dependent manner, but the binding to the prodomain-containing T-cadherin (the R138A mutant) was significantly stronger. These results indicate that the prodomain-containing T-cadherin has higher sensitivity for detecting the T-cadherin binding activity of adiponectin than does prodomain-absent T-cadherin, as expected. Thus, we used the prodomain-containing T-cadherin-based assay for Tcad-AN measurement hereafter.

Our previous findings showed the presence of atherogenic LDL-adiponectin complexes in the atherogenic subfraction of LDL, called L5, in human plasma ²⁸⁾ . Therefore, here we developed a specific ELISA system to measure MAC ( Fig.2D ) . While modification and/or acquisition of atherogenicity of LDL can occur in various ways, we defined atherogenic LDL as having LOX-1-binding capacity as we previously have ³⁴⁾ . Recombinant LOX-1, which is the modified LDL receptor, was used as the capture protein and for detection, and an anti-human adiponectin polyclonal antibody was used to quantify the MAC. A mixture of the serial dilution of adiponectin (0.125 – 1 µg/mL) and the serial dilution of oxLDL (0.125 – 1 µg/mL) resulted in a synergistic increase in the detection of MAC activity in a dose-dependent manner; these results indicate that MAC activity could be detected by this ELISA system.

Correlation between Serum Adiponectin Quantity Measured by Conventional ELISA Method and Tcad-Adiponectin

We examined the ability of the above method to detect Tcad-adiponectin in serum samples from patients (n = 39) at Shinshu University hospital ( Table 1 ) . We detected Tcad-adiponectin in the serum at 1.3±0.4 µg/mL in these patients.

Table 1. Baseline characteristics of subjects in the analysis of correlations between total-adiponectin, HMW-adiponectin, and Tcad-adiponectin.

Variables
N (males/females)	39 (26/13)
Age (years)	62.4±13.5
Body mass index (kg/m2)	24.0±5.3
Diabetes, n (%)	20 (51.3)
Hypertension, n (%)	21 (53.8)
Hyperuricemia, n (%)	3 (7.7)
Dyslipidemia, n (%)	15 (38.5)
Cardiovascular disease, n (%)	18 (46.2)
Ischemic heart disease, n (%)	4 (10.3)
Fatty liver, n (%)	2 (5.1)
Chronic kidney disease, n (%)	11 (28.2)
Dialysis, n (%)	9 (23.1)
Malignant tumor, n (%)	12 (30.0)
Current smoking, n (%)	1 (2.5)
Current drinking, n (%)	11 (27.5)
LDL cholesterol (mg/dL)	103.4±26.5
HDL cholesterol (mg/dL)	52.9±15.3
Triglycerides (mg/dL)	152.8±97.0
Total-adiponectin (μg/mL)	8.0±5.6
HMW-adiponectin (μg/mL)	6.0±5.6
Tcad-AN (μg/mL)	1.3±0.4
Cystatin C (mg/L)	2.0±2.2

Baseline data are expressed as the mean and standard deviation and were statistically evaluated by using an unpaired t test. Categorical variables are presented as the number (n) of subjects and the frequency (percentage) and were analyzed by using the chi-square test. All statistical tests were two- sided.

Hypertension was defined as systolic blood pressure ＞140 mmHg, diastolic blood pressure ＞90 mmHg, and/or use of anti-hypertensive medication. Diabetes was defined as fasting glucose ＞126 mg/dL, glycated hemoglobin ＞6.5%, and/or medication use for diabetes. Chronic kidney disease was defined as positive for proteinuria and/or an eGFR of ＜60 mL/min per 1.73m². Abbreviations: LDL, low-density lipoprotein; HDL, high-density lipoprotein; HMW, high molecular weight; Tcad-AN, T-cadherin bindable adiponectin.

To further verify our detection system, we compared the values we obtained with those from commercially available ELISA kits for HMW-adiponectin and total adiponectin. The concentrations of total adiponectin and HMW-adiponectin were correlated with each other (R² = 0.96, P＜0.0001, Fig.3 , left). In contrast, serum Tcad-adiponectin activity showed a significant but weaker correlation with total serum adiponectin (R² = 0.36, P＜0.0001, Fig.3 , middle) and HMW-adiponectin levels (R² = 0.29, P = 0.0004, Fig.3 , right).

Fig.3. Correlation between serum adiponectin (AN) levels measured by conventional ELISA method and Tcad-AN.

Correlation between total-AN and high-molecular weight (HMW)-AN (left), total-AN and T-cadherin bindable adiponectin (Tcad-AN) (middle), HMW-AN and Tcad-AN (right).

Measurement of MAC and Its Clinical Significance

We quantified the levels of MAC in the serum and assessed their clinical significance in patients undergoing dialysis (hemodialysis patients) compared with healthy controls (n = 35 for each group, age- and sex-matched). Of the 70 samples, MAC was less than the detectable limit in 32, and the value of MAC was treated as 0 in these patients. The average MAC level was 182.6±330.5 ng AN/mL, and the median value was 9.5 ng AN/mL (n = 70).

The number of patients with diabetes or hypertension was higher in the hemodialysis group than in controls ( Table 2 ) . Total cholesterol, LDL cholesterol, and HDL cholesterol serum concentrations were lower in the hemodialysis group than in controls, but serum triglyceride levels were higher in hemodialysis patients.

Table 2. Clinical characteristics of hemodialysis patients and control subjects.

Variables	Control	Hemodialysis	P
n (males/females)	35 (23/12)	35 (23/12)	1
Age (years)	68.2±12.1	68.2±12.1	1
Body mass index (kg/m2)	21.8±2.7	22.5±4.6	0.4224
Diabetes, n (%)	0 (0)	16 (45.7)	＜0.0001
Hypertension, n (%)	11 (31.4)	28 (80.0)	＜0.0001
eGFR (mL/min/1.73m2)	64.7±12.5	6.3±2.6	＜0.0001
Total cholesterol (mg/dL)	198.1±25.1	153.3±46.4	＜0.0001
LDL cholesterol (mg/dL)	113.2±22.0	80.6±32.1	＜0.0001
HDL cholesterol (mg/dL)	74.1±16.6	45.4±17.5	＜0.0001
Triglycerides (mg/dL)	76.9±21.4	107.2±48.1	0.0011

Baseline data are expressed as the mean±standard deviation and were statistically evaluated by using an unpaired t test. Categorical variables are presented as the number (n) of subjects and the frequency (percentage) and were analyzed by using the chi-square test. All statistical tests were two-sided.

Diabetes was defined as fasting glucose ≥ 126 mg/dL, glycated hemoglobin ≥ 6.5%, and/or medication use for diabetes. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, and/or use of anti-hypertensive medication. Abbreviations: LDL, low-density lipoprotein; HDL, high-density lipoprotein.

The hemodialysis group had higher levels of serum total adiponectin and Tcad-adiponectin than did controls ( Fig.4A ) . Levels of lectin-like oxidized low-density lipoprotein receptor-1 ligand containing apoB (LAB) were higher in hemodialysis patients than in controls (98.1±75.6 ng/mL vs.66.3±22.4 ng/mL, respectively; P = 0.0197); oxLDL concentrations did not differ between the groups ( Fig.4A ) . Furthermore, the hemodialysis group had a higher proportion of MAC-high patients than did the control group (62.9% vs. 37.1%, respectively; P = 0.0314, Fig.4B ).

Fig.4. Comparison of serum adiponectin (AN), Tcad-AN, oxLDL, LAB, and MAC levels between hemodialysis patients and control subjects.

(A) Serum levels of AN, T-cadherin bindable adiponectin (Tcad-AN), oxidized LDL (oxLDL) and LOX-1-ligand containing apoB (LAB) in patients on hemodialysis (HD; n = 35) and control subjects (n = 35). The error bars represent the mean±SD of the values obtained.

(B) Comparison of the proportion of MAC-high between HD and controls. MAC-high and -low groups were divided by the median of the MAC values of 70 subjects.

We compared LAB, oxLDL, and adiponectin levels in MAC-high and MAC-low groups in the hemodialysis group. LAB was significantly lower in the MAC-high group than in MAC-low group, whereas oxLDL, total adiponectin, and Tcad-adiponectin levels did not differ ( Fig.5 ) . Since LAB is an index of the biological activity of modified LDL, our findings may suggest that adiponectin interferes with modified LDL activity via MAC formation.

Fig.5. Comparison of oxLDL, LAB, adiponectin (AN) and Tcad-AN levels in hemodialysis patients belonging to MAC-low (open circles, n = 13) and MAC-high (closed circles, n = 22) groups.

The error bars represent mean±SD.

Discussion

Significance of the Modified LDL-Adiponectin Complex

Adiponectin has various anti-diabetic and anti-atherogenic actions. Since LOX-1 mediates the proatherogenic effects of modified LDL on vascular endothelial function, it is reasonable to assume that adiponectin and the oxLDL receptor LOX-1 may have functionally contrasting actions on endothelial function ^{38 , 39)} .

We previously showed that adiponectin forms a complex with oxLDL. In experiments with cultured cells ²⁸⁾ , adiponectin inhibited the uptake of oxLDL into cells expressing LOX-1, cells expressing Scavenger receptor class A (SR-A), endothelial cells, and macrophages, but adiponectin did not inhibit the uptake of LDL into cells via the low-density lipoprotein receptor (LDLR). Adiponectin also suppressed ERK phosphorylation, NF-κB activation, and expression of monocyte chemoattractant protein-1, intercellular adhesion molecule-1, and endothelin-1 in vascular endothelial cells induced by oxLDL.

Here, we have extended those observations and verified the significance of oxLDL-adiponectin complex formation at the vascular tissue level. As expected, as seen with neutralizing anti-LOX-1 antibody, physiological concentrations of adiponectin suppressed the vasodilatory dysfunction caused by oxLDL. Furthermore, similar results were obtained with the simultaneous administration of anti-LOX-1 antibody and adiponectin and the administration of adiponectin alone, suggesting that the site of action is the same. Thus, we showed that the formation of the oxLDL-adiponectin complex protects the vascular endothelium from oxLDL-induced vasorelaxation dysfunction.

Novel Method for Quantifying the Modified LDL-Adiponectin Complex (MAC)

To study the clinical significance of the complex formation between adiponectin and oxLDL, we have developed the first measurement system for quantifying MAC activity. Using this method, we verified that modified LDL and adiponectin form a complex in the blood in measurable amounts.

In patients on hemodialysis, the percentage of patients falling into the MAC-high group was significantly higher than that seen in healthy controls. Since the concentrations of total adiponectin in the hemodialysis group were higher with similar concentrations of oxLDL compared with control, MAC formation could more easily occur in the hemodialysis group.

The level of biologically active modified LDL, LAB, was higher in the hemodialysis group, despite the possible neutralization of modified LDL by MAC formation in the hemodialysis group. However, on further analysis by dividing the hemodialysis group into two groups based on high and low MAC levels, LAB was significantly lower in the MAC-high group than in the MAC-low group. This suggests that MAC formation in humans may inhibit the binding of modified LDL to its receptors, including LOX-1. We found no significant difference between the two groups in the concentration of oxLDL or adiponectin as measured using antibodies. Thus, it is unlikely that the difference we found in LAB was a secondary result caused by a difference in the concentration of oxLDL or adiponectin.

Regulation of Biological Activity of Adiponectin by Its Binding Proteins

Adiponectin-binding proteins include various chemokines, cystatin C ⁴⁰⁾ , E-selectin ligand-1, and Mac-2 binding protein ⁴¹⁾ . Of these, cystatin C reportedly binds to adiponectin and suppresses the anti-inflammatory effects and reduces the clearance of adiponectin in vivo ⁴²⁾ . Moreover, the concentration of the cystatin C-adiponectin complex in the blood has been correlated with the instability of coronary artery plaques ⁴³⁾ . Thus, the blood levels of MAC may also be related to the progression and prognosis of coronary artery disease, but this assumption warrants further study.

The modulation of adiponectin activity by adiponectin-binding proteins has been recently recognized. Thus, in this study, we developed a new system for measuring the physiological activity rather than the absolute amount of adiponectin binding activity to T-cadherin, an adiponectin receptor. The adiponectin concentration in the blood as measured by commercially available kits for total and HMW adiponectin showed a strong correlation with each other, whereas adiponectin activity measured by T-cadherin binding and adiponectin concentration measured by the above kits showed a weaker correlation. This may suggest that the binding of adiponectin to T-cadherin is affected by substances in the blood that can modulate adiponectin activity, such as modified LDL (as we have shown here and previously) ²⁸⁾ .

Lipoproteins That Form Complexes with Adiponectin in vivo

Using LDL oxidatively modified in vitro as a model lipoprotein for modified LDL, we showed that it forms a complex with adiponectin. However, the LDL complexed with adiponectin detected in the blood is not necessarily the same kind of modified LDL. There are various modes of LDL modification, and each of the proteins, phospholipids, cholesterols, and carbohydrate chains in LDL could undergo modification, including oxidation, hypochlorous acid modification, carbamylation, glycation, and desialylation ^{44 - 50)} .

Although the molecule is simply called oxidized or modified LDL, it is a group of heterogeneous lipoproteins that have undergone various modifications. Additionally, dietary lipoproteins such as remnants, which exhibit properties similar to modified LDL, and lipoproteins such as the negatively charged LDL subfraction (L5) show similar atherogenic activity to oxLDL ^{32 , 51)} . Indeed, we have previously shown that adiponectin is concentrated in the L5 fraction ²⁸⁾ . Furthermore, the possibility of other modified LDL or remnants cannot be excluded from lipoproteins that form a complex with adiponectin in vivo.

Limitation of the Study

The method for measuring MAC activity developed in this study detects the binding activity to the modified LDL receptor LOX-1. The MAC formation suppresses the binding of modified LDL to LOX-1, whereas the complex measured in this study still has LOX-1 binding activity. Measuring the complex that has completely lost LOX-1-binding activity may be more useful as a biomarker. Furthermore, the number of samples used in the human analysis was limited. In the future, we plan to perform similar studies in larger human cohorts and a longitudinal study on the relationship between MAC concentration and atherosclerosis progression, resulting in ischemic cardiovascular diseases.

Conclusion

Using a novel assay, we confirmed the presence of measurable levels of MAC in human serum. This complex formation neutralizes the effects of modified LDL on blood vessels, but whether the MAC formation in human blood affects vascular function and the progression of arteriosclerosis is unknown.

Acknowledgements

The authors thank Rebecca Bartow, PhD, of the Department of Scientific Publications at The Texas Heart Institute for editorial contributions.

Conflicts of Interest

T.S. and A.K. are the inventors and have patents pending on the adiponectin ELISA methods.

Notice of grant support: This study was supported by the Japan Society for the Promotion of Science, JSPS KAKENHI (Grant No. 22K19707 and 21H04854 to T.S., No. 16K19192 to A.K.), and Shinshu University POC fund to T.S.

Contribution statement: T.S. conceptualized the study and wrote the manuscript. A.K. and T.S designed experiments. M.S. and A.K. performed the experiments and wrote the draft of the manuscript. Y.Y. analyzed the data and wrote the draft of the manuscript. R.V., F.S, and S.T.D. performed animal experiments. S.J., H.H. and T.U. assisted in preparation of the experiments and commented on the manuscript. C.C. commented on the draft of the manuscript. All authors have contributed to data interpretation and approved the final draft of the manuscript.