High levels of IgM against methylglyoxal-modified apolipoprotein B100 are associated with less coronary artery calcification in patients with type 2 diabetes


Daniel Engelbertsen, CRC Entrance 72; 91:12, SUS, Malmö, 20502 Malmö, Sweden.
(fax: +46 40391212; e-mail: daniel.engelbertsen@med.lu.se).


Abstract.  Engelbertsen D, Anand DV, Fredrikson GN, Hopkins D, Corder R, Shah PK, Lahiri A, Nilsson J, Bengtsson E (Skåne University Hospital, Lund University, Malmö, Sweden; Cardiac Imaging and Research Centre, Wellington Hospital, London, UK; Malmö University, Malmö, Sweden; King’s College Hospital, London, UK; William Harvey Research Institute, London, UK; Cedars-Sinai Medical Center, Los Angeles, CA, USA; David Geffen School of Medicine, UCLA School of Medicine, Los Angeles, CA, USA). High levels of IgM against methylglyoxal-modified apolipoprotein B100 is associated with less coronary artery calcification in patients with type 2 diabetes. J Intern Med 2012; 271: 82–89.

Objective.  Advanced glycation end products (AGE) have been implicated in diabetic vascular complications through activation of pro-inflammatory genes. AGE-modified proteins are also targeted by the immune system resulting in the generation of AGE-specific autoantibodies, but the association of these immune responses with diabetic vasculopathy remains to be fully elucidated. The aim of this study was to determine whether antibodies against apolipoprotein B100 modified by methylglyoxal (MGO-apoB100) are associated with coronary atherosclerosis in patients with type 2 diabetes.

Methods.  We measured antibodies against MGO-apoB100 in plasma from 497 type 2 diabetic patients without clinical signs of cardiovascular disease. Severity of coronary disease was assessed as coronary artery calcium (CAC) imaging. Immunoglobulin (Ig)M and IgG levels recognizing MGO-apoB100 were determined by enzyme-linked immunosorbent assay.

Results.  Anti-MGO-apoB100 IgM antibody levels were higher in subjects with a low to moderate CAC score (≤400 Agatston units) than in subjects with a high score (>400 Agatston units; 136.8 ± 4.4 vs. 101.6 ± 7.4 arbitrary units (AU), < 0.0001) and in subjects demonstrating no progression of CAC during 30 months of follow-up (136.4 ± 5.7 vs. 113.9 ± 6.2 AU in subjects with progression, < 0.0001). Subjects with a family history of premature myocardial infarction had lower levels of anti-MGO-apoB100 IgM. Female subjects had higher levels of anti-MGO-apoB100 antibodies and lower CAC than men. Accordingly, high levels of IgM against MGO-apoB100 are associated with less severe and a lower risk of progression of coronary disease in subjects with type 2 diabetes.

Conclusions.  Although conclusions regarding causal relationships based on epidemiological observations need to be made with caution, our findings suggest the possibility that anti-MGO-apoB100 IgM may be protective in diabetic vasculopathy.


advanced glycation end product


apolipoprotein B100


coronary artery calcium


low-density lipoprotein




Coronary artery disease (CAD) is a major cause of mortality in patients with type 2 diabetes [1]. The mechanisms through which type 2 diabetes contributes to a more aggressive development of atherosclerosis remain to be fully understood, but considerable attention has focussed on the possible role of glucose-derived advanced glycation end products (AGE). The formation of AGE on nucleophilic side chain residues of amino acids (e.g. lysine and arginine side chains) is accelerated by hyperglycaemia [2], and animal studies have shown that the presence of AGE epitopes in atherosclerotic plaques is increased by diabetes [3]. AGE act as ligands for specific cell surface receptors (receptor for AGE; RAGE), and activation of RAGE is associated with enhanced expression of several NF-κB-regulated pro-inflammatory genes [4]. Inhibition of RAGE has been shown to suppress arterial inflammation and development of atherosclerosis in diabetic apolipoprotein E-null mice providing direct evidence for a role of AGE in diabetic vascular complications [5].

Advanced glycation end product–modified structures are also targeted by the immune system, and antibodies against various AGE-modified proteins have been observed in patients with diabetes [6, 7]. Whether immune responses against AGE epitopes in modified self-antigens are pathogenic and contribute to the development of vascular disease or have a protective role remains to be clarified. The role of immune responses against modified self-antigens, particularly oxidized low-density lipoprotein (LDL), in atherosclerosis has attracted increasing attention during recent years [8, 9]. Experimental studies have provided conclusive evidence that Th1-type immunity against self-antigens modified by hypercholesterolaemia is an important driving force in atherosclerosis [10]. However, these studies have also revealed the existence of a protective immunity largely mediated by antibodies and regulatory T cells [11]. Several lines of evidence have suggested that so-called natural phosphorylcholine-specific immunoglobulin (Ig)M antibodies have a protective role [12] and a similar function has been attributed to IgG against certain peptide sequences in apolipoprotein B100 (apoB100) [13]. We have recently shown that high IgG levels against two such sequences, apoB p45 and p210, are associated with less coronary disease in patients with type 2 diabetes [14].

Reactive α-dicarbonyls such as methylglyoxal (MGO) are, along with glucose, key producers of AGE. MGO, which is elevated in individuals with type 2 diabetes [15], is generated by autooxidation of glucose, fragmentation of triosephosphates and catabolism of ketone bodies [2, 16]. Accumulation of an AGE adduct generated by MGO modification of the lysine residue N-ε-carboxyethyl-lysine (CEL) has been shown to be increased in plaques of diabetic hyperlipidaemic apoE-null mice [17], and CEL-modified LDL has been detected in immune complexes isolated from patients with diabetes [7]. To study the possible role of immune responses against MGO-modified self-antigens in diabetic macrovascular complications, we determined the relation between IgG and IgM autoantibody levels against MGO-apoB100 and the severity of coronary atherosclerosis as assessed by the degree of coronary calcification in a cohort of 497 subjects with type 2 diabetes.

Materials and methods

MGO-apoB100 antibody enzyme-linked immunosorbent assay (ELISA)

Methylglyoxal-apoB100 was generated by co-incubation of apoB100 (Calbiochem, La Jolla, CA, USA) with 100 mmol L−1 MGO (Sigma, St. Louis, MO, USA) in 0.2 mol L−1 phosphate-buffered saline (PBS) at 37 °C for 24 h. The protein was subsequently dialysed against 0.15 mol L−1 PBS (pH 7.4) before storage at −20 °C. AGE modification was verified by demonstrating a 32-fold increase in AGE-specific fluorescence (excitation 370 nm, emission 440 nm) of MGO-apoB100 relative to native protein and by argpyrimidine fluorescence (excitation 320 nm, emission 380 nm) demonstrating an increase in MGO-apoB100 relative to native protein (106 vs. 0 fluorescence units). CEL epitopes on MGO-apoB100 were measured by ELISA using a monoclonal anti-CEL antibody (KNH-30; Cosmo Bio, Tokyo, Japan). Briefly, MGO-apoB100 was coated onto microtitre plates, and then wells were blocked with Superblock (ThermoScientific, Rockford, IL, USA) and washed, and anti-CEL followed by biotinylated anti-mouse IgG (Vector Laboratories, Burlingame, CA, USA) and alkaline phosphatase-conjugated streptavidin was added. AGE modifications were also assayed using anti-MGO-AGE, primarily recognizing argpyrimidine (antibodies-online, clone 6B) in ELISA as described earlier.

An ELISA measuring IgM and IgG in plasma against MGO-apoB100 was performed essentially according to Fredrikson et al. [18], except for coating with 10 μg mL−1 protein and the use of streptavidin–alkaline phosphatase from Biolegend (San Diego, CA, USA). Absorbance values were normalized to a control plasma pool (pooled from 11 individuals without type 2 diabetes). The intra-assay coefficient of variation was 11.7% for IgG and 3.6% for IgM, and the inter-assay coefficient of variation was 14% and 16% for IgG and IgM, respectively. Antigen specificity was determined by a soluble-phase competitive ELISA. In brief, native and MGO-modified proteins (apoB100, bovine serum albumin (BSA), human serum albumin or human plasma fibronectin) were added to control plasma (diluted 1 : 5). After 2-h incubation at room temperature followed by overnight incubation at 4 °C, immune complexes were removed by centrifugation at 1000 g. Subsequently, the plasma supernatant was collected, diluted to 1 : 100, added to wells coated with MGO-apoB100 and analysed as above. Antibodies recognizing glyoxal-modified apoB100 were measured in five patients, and the levels were similar to those of anti-MGO-apoB100.

Analysis of AGE proteins in plasma

An ELISA was developed to measure AGE-modified apoB100 in plasma. Wells were coated with polyclonal anti-apoB100 (Abcam, ab 20737-1) and blocked with Superblock. Wells were then incubated with test plasma overnight. Bound AGE-apoB100 was detected with mouse anti-CEL (KNH-30; Cosmo Bio), followed by biotinylated goat anti-mouse IgG (Vector Laboratories) and alkaline phosphatase-conjugated streptavidin. To measure total AGE-modified proteins in plasma, wells were coated with test plasma overnight and blocked with Superblock, and AGE proteins were detected with anti-CEL as described above. In both cases, the amount of AGE proteins was below the detection limit of our assays.

Study population

Plasma samples were derived from 510 type 2 diabetic individuals without pre-existing cardiovascular disease. Thirteen patients were excluded from the analysis because of sample unavailability. Inclusion and exclusion criteria have been described elsewhere [19]. The study was approved by the local research ethics committees of the participating institutions, and all subjects gave informed consent.

Coronary artery calcium imaging

Coronary artery calcium (CAC) imaging was performed using electron beam computed tomography, and coronary calcium levels are given in Agatston units. The CAC imaging protocol [19] and CAC stratification [14, 20] have been previously described. Determination of CAC progression has been previously defined [14].

Statistical analysis

Differences in baseline characteristics were tested using χ2- or t-tests, as applicable. Skewed variables were log-transformed before statistical analysis. Spearman’s and Pearson’s correlation coefficients were used, as appropriate, to examine associations between continuous variables. Linear regression analysis was used to correct for interferences.  0.05 was considered significant.


Methylglyoxal modification was verified by demonstrating relative increases in both AGE-dependent (370/440 nm) and argpyrimidine-dependent fluorescence (320/380 nm) and antibodies recognizing CEL and argpyrimidine epitopes (Fig. 1a, b) in MGO-apoB100 compared to native apoB100. Next, we determined the specificity of our anti-MGO-apoB100 ELISA. The binding of pooled control plasma to MGO-apoB100-coated ELISA plates was reduced by pre-incubation of the plasma with MGO-apoB100, but not by pre-incubation with native apoB100, MGO-modified BSA, native BSA (Fig. 1c,d) or native and MGO-modified human serum albumin and human plasma fibronectin (Fig. S1), demonstrating the specificity of the assay.

Figure 1.

Anti-advanced glycation end products ELISA and specificity assessment of anti-methylglyoxal (MGO)-apolipoprotein B100 (apoB100) plasma antibodies. N-ε-carboxyethyl-lysine (CEL) epitopes (a) and argpyrimidine (b) on MGO-modified apolipoprotein B100 (MGO-apoB100) and apolipoprotein B 100 (apoB100) were measured by ELISA. (c, d) Comparison of anti-MGO-apoB100 signal inhibition by pre-incubation of pooled plasma from type 2 diabetic subjects (n = 11) with designated concentrations of apoB100, MGO-apoB100, BSA or MGO-modified BSA. Values are given as the fraction of the absorbance of plasma pre-incubated with protein (Abs inhibited) divided by the absorbance of control plasma (Abs control), for IgM (c) and IgG (d). Each data point (a–d) represents the mean of triplicate wells ± SD.

The subjects included in the present analysis were recruited from a cohort of patients with type 2 diabetes participating in a study of the ability of coronary calcium score to predict silent myocardial ischaemia and short-term cardiovascular events. The characteristics of the study group are described in Table 1. All subjects were free of clinical manifestations of CAD at the time of investigation.

Table 1. Characteristics of the study group
  1. HDL, high-density lipoprotein; LDL, low-density lipoprotein.

  2. Values are expressed as mean ± standard deviation or number of individuals (percentage of the total group).

Age52.7 ± 8.4
Sex195 (39%) women/302 (61%) men
Duration of diabetes (years)8.1 ± 6.0
HbA1c (%)8.2 ± 1.7
BMI (kg m−2)28.5 ± 5.0
Systolic blood pressure (mmHg)137 ± 16
Diastolic blood pressure (mmHg)84 ± 12
Total cholesterol (mmol L−1)4.8 ± 0.9
LDL cholesterol (mmol L−1)2.7 ± 0.8
HDL cholesterol (mmol L−1)1.3 ± 0.4
Triacylglycerol (mmol L−1)1.9 ± 1.1
C-reactive protein (mg L−1)8.1 ± 30
Current smokers 96 (19%)
Insulin therapy105 (21%)
Statin therapy195 (39%)

Both IgG and IgM against MGO-apoB100 could be detected in almost all subjects (Fig. 2). Anti-MGO-apoB100 IgM antibody levels were higher in subjects with a low to moderate CAC score (≤400 Agatston units) than in subjects with a high CAC score (>400 Agatston units; 136.8 ± 4.4 vs. 101.6 ± 7.4 arbitrary units (AU), < 0.0001, Fig. 3a). Accordingly, there was a significant inverse association between anti-MGO-apoB100 IgM levels and coronary calcification expressed as log CAC (r = −0.164, < 0.0001) (Fig. S2). This association remained significant after controlling for age, body mass index, hypertension, systolic blood pressure, duration of diabetes, HbA1c, LDL cholesterol, high-density lipoprotein (HDL) cholesterol and triacylglycerol using a linear regression model. Follow-up CAC imaging was performed in 398 patients after a mean follow-up time of 2.5 ± 0.4 years. Patients with no progression of coronary calcification had higher levels of anti-MGO-apoB100 IgM antibodies than patients demonstrating progression of coronary calcification (136.5 ± 5.8 vs. 113.9 ± 6.2 AU, = 0.008; Fig. 3b). Patients reporting a family history of premature CAD had lower levels of IgM anti-MGO-apoB100 than subjects with no history of CAD (118.0 ± 7.0 vs. 139.3 ± 4.9 AU, = 0.014). There was no association between the levels of anti-MGO-apoB100 IgG and either severity or progression of CAC (data not shown).

Figure 2.

Levels of anti-methylglyoxal apolipoprotein B100 in patients with type 2 diabetes. Plasma IgG and IgM antibody levels were measured by ELISA. Absorbance level of each individual sample (n = 497) is expressed as percentage of nondiabetic control plasma (n = 11).

Figure 3.

Association between anti-methylglyoxal (MGO)-apolipoprotein B100 (apoB100) IgM antibodies and coronary artery calcium score (CAC) in type 2 diabetic subjects. Levels of anti-MGO-apoB100 IgM antibodies in patients with low to moderate CAC (≤400 Agatston units) compared with patients with severe to extensive CAC (>400 Agatston units) (a). Anti-MGO-apoB100 levels in patients with progression of calcification compared to levels in nonprogressors (b). The values are mean ± SEM. Antibody levels are given as percentage of control.

The level of anti-MGO-apoB100 IgM in plasma correlated with age (r = −0.165, < 0.0001) and LDL (r = −0.094, < 0.05), but neither IgM nor IgG correlated with systolic blood pressure, HDL, triglycerides or HbA1c. Both IgM and IgG anti-MGO-apoB100 antibody levels were associated with the duration of diabetes (r = −0.109, P = 0.015 and r = −0.106, P = 0.018, respectively). Female subjects had significantly higher levels of anti-MGO-apoB100 IgM than male subjects (147.5 ± 6.8 vs. 123.0 ± 4.9, =0.003). When gender was included as a variable in the linear regression model, the association between IgM anti-MGO-apoB100 antibodies and log CAC was no longer significant (= 0.079). Analysing male and female subjects separately revealed that the correlation between IgM and log CAC was significant amongst men (r = −0.173, = 0.002; n = 302), but not amongst women (r = −0.071, = ns; n = 195). Women were also characterized by a lower CAC score than men (95.7 ± 423.6 vs. 278.0 ± 680.9 Agatston units, < 0.001). There was no difference in IgM or IgG antibody levels between pre- and postmenopausal women. Postmenopausal women (n = 104) had higher CAC than premenopausal women (n = 91) (160 AU ± 57 vs. 21 AU ± 77, < 0.0001), whereas there was no difference in progression of the disease.

We did not find any evidence for significant co-variations between IgM or IgG anti-MGO-apoB100 antibody levels and neuropathy, retinopathy or nephropathy. In addition, there were no significant correlations between microalbuminuria, levels of urea or creatinine and IgM or IgG anti-MGO-apoB100.


The results of the present study suggest that patients with type 2 diabetes who have high plasma levels of IgM against the MGO-modified LDL-binding protein apoB100 are characterized by less-severe coronary disease and a lower risk of coronary disease progression. Although conclusions regarding causal relationships based on epidemiological associations need to be made with caution, our findings suggest a possible protective role of anti-MGO-apoB100 IgM in diabetic vasculopathy. This notion is in line with previous studies suggesting a protective role of IgM in atherosclerosis by recognizing modified phospholipids in oxidized LDL [21, 22] and by binding to oxidation-specific epitopes on apoptotic cells [12, 23]. The mechanisms through which these IgM antibodies exert their atheroprotective action are not fully understood but may include facilitation of the removal of damaged cells and lipoproteins. It is an interesting possibility that IgM recognizing AGE-modified self-antigens may have a similar function either by facilitating the removal of these antigens or by blocking the binding of AGE epitopes to RAGE.

The first evidence for the existence of autoantibodies against AGE-modified self-antigens came from Witztum et al. [24] who demonstrated that plasma from patients with diabetes could react with glycated LDL and albumin. Shibayama and colleagues [6] subsequently confirmed this finding and identified the N-ε-carboxymethyl-lysine (CML) epitope as a major target for these autoantibodies. They also reported increased levels of autoantibodies against CML-modified albumin in patients with diabetic nephropathy. In the only previous study addressing the association between AGE autoantibodies and cardiovascular disease in diabetes, Lopes-Virella et al. [25] found no relation between the level of anti-AGE-modified LDL antibodies and progression of carotid intima-media thickness amongst 1229 patients with type 1 diabetes participating in a follow-up study of the Diabetes Control and Complications Trial. Several factors may have contributed to the different outcomes of the present study and that of Lopes-Virella and colleagues. Whereas the latter study included only patients with type 1 diabetes, the present study was restricted to patients with type 2 diabetes. Accordingly, insulin sensitivity and other metabolic risk factors differ markedly. The epitope specificity of the ELISAs used to determine AGE antibodies was also not the same. Even though the spectra of AGE epitopes formed by glucose and MGO overlap, as MGO is generated during several steps of classical AGE formation by glucose [16], the frequency of MGO-generated epitopes would be greatly enhanced on proteins modified by MGO compared to those modified by glucose. Although LDL and apoB100 are comparable antigens, modification of LDL may generate additional epitopes compared to those formed on apoB100 alone. Finally, the severity of atherosclerosis was assessed by two different techniques at two different locations (i.e. the coronary vs. the carotid arteries) in the two studies.

The level of anti-MGO-apoB100 IgM was found to be significantly higher in female than in male subjects. Women were also characterized by less-severe coronary disease as assessed by the CAC score. Although the association between anti-MGO-apoB100 IgM and CAC was significant only amongst men, the general trend was the same in both groups. It may be that the lower and more narrowly distributed CAC values in the female group make it difficult to detect an association between anti-MGO-apoB100 IgM and CAC in women. In another study, Su et al. [22] analysed antibody levels against oxidized LDL and their correlation with atherosclerosis in patients with hypertension. It is interesting that they found higher levels of IgM against oxidized LDL in women than in men, and women also had less plaque.

Of note, we observed an association between family history of premature CAD and low anti-MGO-apoB100 IgM antibody levels, suggesting the possibility of genetic factors in the regulation of these antibodies.

In this study, we found no correlation between IgG antibodies against MGO-apoB100 and CAC. It is interesting to note that several studies have found discrepancies between IgM and IgG antibodies against oxidized LDL and an association with cardiovascular disease [21, 22, 26]. This isotype difference may have several explanations. One possibility could be that IgM antibodies are T-cell independent, whereas most IgG subtypes are T-cell dependent. T cells produce cytokines, which could have effects on plaque formation. In addition, studies in mice support a protective role of IgM in cardiovascular disease; Ldlr−/− mice deficient in serum IgM showed a 7-fold increase in plaque area compared with Ldlr−/−control mice [27].


In summary, we have demonstrated that high levels of IgM antibodies against MGO-apoB100 are associated with a less-severe coronary disease in patients with type 2 diabetes. These observations provide evidence for a role of immune response against AGE-modified self-antigens in diabetic vascular complications and suggest that IgM recognizing such structures may have a protective function.

Conflict of interest

The authors declare that there is no conflict of interest associated with this study.


We are grateful to Margaretha Persson Ph.D. for providing a control plasma pool from diabetic patients. This study was supported by grants from the Swedish Medical Research Council, the Swedish Heart-Lung Foundation, the Söderberg Foundation, the Albert Påhlsson Foundation, the Knut and Alice Wallenberg Foundation, the Malmö University Hospital Foundation, the Åke Wiberg Foundation, the Royal Physiographic Society Lund, the Magnus Bergvall Foundation, the Lundström Foundation, the Crafoord Foundation, VINNOVA, the Foundation for Strategic Research (SSF), the Harrow Cardiovascular Research Trust, the Michael Tabor Foundation, GE Healthcare Ltd, Bristol Myers Squibb Medical Imaging Inc, the Derrick Smith Research Grant and the European Community’s Sixth Framework Programme contract (‘IMMUNATH’) LSHM-CT-2006-037400.