Introduction

Diabetes affects more than 171 million people worldwide [1]. Type 1 diabetes occurs when an individual produces very little or no insulin. Type 2 diabetes results when the body is unable to properly use the insulin it produces, in combination with a relative insulin deficiency [1]. Individuals with both Type 1 and Type 2 diabetes exhibit hyperglycemia and often develop vascular complications, which account for an increase in morbidity and mortality in these individuals. Nephropathy is one of the most prevalent microvascular complications of diabetes affecting 15–25% of Type 1 and 20–40% of Type 2 patients, and is the most common cause of end stage renal failure in the western world [2]. Features of nephropathy include persistent albuminuria, a progressive decline in renal function and alterations in kidney morphology. Determining a sensitive predictor or establishing the etiological basis of nephropathy would allow for appropriate intervention with a reduction in morbidity and mortality in individuals with diabetes. At present, the exact cause of this diabetic complication is not clear. Advanced glycation end products (AGEs) are elevated in diabetes, and there is strong evidence that they may be involved in the pathogenesis of its vascular complications [3–10].

In diabetes, AGEs are formed when excess glucose and other reducing sugars react with free amino and sulfhydryl groups of amino acids in proteins, and through a series of reactions and rearrangements form irreversible conjugates that alter protein structure and function. Glyoxal and methylglyoxal, aldehydes produced from enzymatic and non-enzymatic glucose and lipid metabolism [11, 12], are elevated in diabetes [13] and also modify amino acids to form AGEs [14] (Fig. 1). Aldehydes are up to 20,000 times more reactive than glucose in glycation processes [15]. Once formed, AGEs also act indirectly through cell surface receptors such as receptors of AGES (RAGEs) and scavenger receptors to cause alterations in protein function [2, 16].

Fig. 1
figure 1

Schematic diagram showing the formation of advanced glycation end products from glucose, or from the aldehydes, methylglyoxal and glyoxal

Full size image

Several studies measuring non-specific AGEs have shown them to be elevated in diabetes and associated with nephropathy [4, 5, 10]. A number of specific AGEs derived from modification to lysine and arginine, have been identified including carboxymethyl lysine (CML), carboxyethyl lysine (CEL), and pentosidine, and there is some evidence that they contribute to nephropathic kidney changes [3, 6–9, 17, 18]. However, other AGEs, particularly those formed with cysteine residues of proteins, may also be involved in the pathology of diabetic nephropathy.

Free sulfhydryl groups of cysteine are more reactive than free amino groups of lysine or arginine [19], and are critical to the normal function of many cellular proteins including enzymes such as nitric oxide synthase (NOS) [20], glyceraldehyde-3-phosphate dehydrogenase (GAPDH) [21], glutathione peroxidase [22], Ca2+-ATPase [23], and tyrosine phosphatases [24], and membrane vascular calcium channels [25] and receptors [26]. Thus, the ubiquitous and functional nature of these cysteine sulfhydryls makes them a likely target for AGE formation with the potential for widespread impact. Modification to protein sulfhydryl groups has been shown to cause changes typical of diabetic vascular disease including increased oxidative stress [22, 27], endothelial dysfunction [20], altered calcium handling [23], and changes to cell signaling [24]. However, information regarding levels of specific cysteine AGEs and their contribution to nephropathy in diabetes is scant [28–30].

The AGEs carboxymethyl cysteine (CMC) and carboxyethyl cysteine (CEC) are formed by the binding of glyoxal and methylgyoxal, respectively, to cysteine sufhydryl groups of proteins [19] (Fig. 2). We could not find any studies reporting the isolation or measurement, of CEC. To the best of our knowledge, plasma protein CEC and CMC have not been quantified in individuals with diabetes, nor is there any evidence concerning their relationship to complications of this condition. Therefore, the objective of this study was to develop a method to measure the cysteine AGEs, CMC and CEC, assess their levels in a group of patients with diabetes and in controls without diabetes, and determine if a relationship exists between these AGEs and the albumin creatinine ratio (ACR), an accepted indicator of diabetic nephropathy.

Fig. 2
figure 2

The aldehydes, glyoxal, and methylglyoxal bind with SH groups of amino acids in proteins to form the advanced glycation end products carboxymethyl cysteine (CMC) and carboxyethyl cysteine (CEC), respectively

Full size image

Materials and methods

Materials

CMC, cysteine and bromopropionic acid were purchased from Sigma-Aldrich Canada (Ontario, Canada). Methanol was of HPLC grade and other chemicals were of analytical grade, and were also purchased from Sigma-Aldrich Canada. l-methionine (Methyl-D3, 98%) was obtained from Cambridge Isotope Laboratories (Massachusetts, USA).

Subjects

The Human Investigation Committee of Memorial University of Newfoundland granted the ethics approval for the study. Forty-two patients with Diabetes Mellitus (DM group) including both Type 1 and Type 2 between the ages of 18 and 36 years were selected from the patient population who had blood collected as part of their diabetes care program. Eighteen patients without diabetes (ND group), aged 17–35 years, were selected based on a normal hemoglobin A1C (HbA1C), from a patient population who had EDTA blood collected for measurement of complete blood count as part of their routine health care. Patient consent was not required to harvest samples for study analysis.

Blood and urine biochemistry

As the control subjects were selected from patients undergoing routine EDTA blood collection for other purposes, urine and serum specimens were not available. Therefore, urine albumin and creatinine, and serum creatinine were not measured in this group. For all study subjects, HbA1C was measured on EDTA blood using a Tosoh G7 Automated HPLC (Tosoh Bioscience Inc., USA). Plasma isolated from the same blood sample was used for measurement of CMC and CEC. All subjects with diabetes had data on serum creatinine and ACR. Measurement of serum and urine creatinine was done using a Synchron LX® system (Beckman Coulter, USA) which determines creatinine concentration by means of the Jaffe rate method. Urine albumin was measured by nephelometry using an Immage® Immunochemistry system (Beckman Coulter, USA). ACR was calculated (urine albumin (mg)/urine creatinine (mmol)) and used as an indicator of nephropathy [31]. The ACR value for each patient in the DM group was the average of a minimum of three measurements, over a 1-year period. An ACR value of >2.0 mg/mmol and >2.8 mg/mmol in males and females, respectively, is indicative of nephropathy.

Synthesis of CEC

CEC was not available for commercial purchase so a method was adapted to synthesize this product. Preparation was based on the procedure of Anacardio et al [32]. Briefly 400 mg of cysteine was dissolved in 7.2 ml water and the pH was adjusted to 7.6 with 1 M NaOH. Bromopropionic acid, 620 mg, was dissolved in 400 μl water and the pH was also adjusted to 7.6. Both solutions were mixed together at room temperature for 12 h and the pH was adjusted to 7.6 using 1 M ammonium hydroxide. The solution was then vacuum dried. The peak corresponding to CEC’s molecular weight plus 1 (M + 1 = 194) was isolated by LC-MS/MS using a reverse phase C18 column (Gemini C18, 5 μm, 250 mm × 10 mm, Phenomenex, California, USA) and was collected.

Preparation of samples

Pooled plasma from patients without diabetes, as well as plasma samples from the DM and ND groups, was prepared based on the procedure of Requena et al. [33]. Briefly 200 μl of EDTA plasma was precipitated with an equal volume of 20% trichloracetic acid (TCA), the pellet was washed with 10% TCA, delipidated once with 2 ml methanol:ether (3:1), and dried under a stream of nitrogen. The samples were hydrolyzed in 6N HCl at 110°C for 24 h. and then dried by vacuum. Pooled plasma hydrolysate from patients without diabetes was used as a matrix in preparation of the standard curve as described next.

Preparation of stock solutions

Stock solutions of CMC and CEC were prepared by dissolving accurately weighed product in 10 M HCl/methanol (1:100, v/v) to give a final concentration of 1 mg/l for CMC and 5 mg/l for CEC. The standard curve was constructed by diluting the stock standard CMC to final concentrations ranging from 0 μg/l to 125 μg/l and CEC from 0 μg/l to 625 μg/l in plasma protein hydrolysate. Calibration curves were corrected for endogenous CMC and CEC in the plasma protein hydrolysate.

Extraction and derivitization procedure

l-methionine (Methyl-D3), 20 μg, was added as an internal standard to all samples (patient samples and standards) followed by 1 ml 10M HCl/methanol (1:10, v/v). The samples were vortex-mixed and then heated to 70°C for 30 min for methylation of the carboxylate groups of CMC, CEC, and the internal standard. The samples were centrifuged at 15,000 RPM for 5 min to remove any particulate and the supernatants were transferred to another tube, ready to be injected onto the LC-MS/MS system.

Measurement of CMC and CEC by LC-MS/MS

All samples were analyzed by LC-MS/MS on the Waters AllianceHT 2795-Micromass Quattro Ultima system using a C18 Column (X Terra® MS C18, 3.5 μm, 2.1 × 50 mm, Waters, Massachusetts, USA). Methylated CMC and CEC were separated isocratically in 0.25% trifluoroacetic acid and methanol (97.5:2.5) at a flow rate of 0.35 ml/min and at ambient temperature. The run time was 8 min. Methylated CMC was determined by multiple reaction monitoring (MRM) 208 > 191, with collision energy set at 6; methylated CEC, MRM 222 > 145, with collision energy set at 10, and methylated internal standard, MRM 167 > 107, with collision energy set at 8. Representative chromatograms of derivatized L-methionine, CMC and CEC are shown in Fig. 3. Using 10 plasma replicates, CMC and CEC were determined to have good reproducibility with this method. Coefficient of variation (calculated as (SD/mean) × 100) was 7.1 for CMC and 3.9 for CEC. The minimum detection limits for CMC and CEC were 2.3 μg and 21.4 μg, respectively.

Fig. 3
figure 3

Representative chromatograms of derivitized (a) internal standard, l-Methionine, monitored by multiple reaction monitor (MRM) 167 > 107; (b) glyoxal-derived plasma protein cysteine advanced glycation end product (AGE), carboxymethyl cysteine (CMC), MRM 208 > 191, 67.6 μg; and (c) methylglyoxal-derived plasma protein cysteine AGE, carboxyethyl cysteine (CEC), MRM 222 > 145, 640.5 μg as measured using a liquid chromatography-mass spectrophotometry method

Full size image

Statistical analysis

Data were analyzed using SPSS for Windows (release 12.0). Results are expressed as mean ± standard deviation. Independent sample t-test was used to compare groups. The relationship between CMC, CEC, HbA1C, serum creatinine, and ACR was analyzed using correlation analysis. Multiple linear regression was used to evaluate the role of HbA1C, serum creatinine, CMC and CEC as predictors of ACR. A P-value less than 0.05 was considered significant.

Results

There was no difference between the age and sex distribution between patients in the DM and ND groups. The mean HbA1C (%) was significantly higher in the DM group (mean ± SD, 8.5 ± 1.6) as compared to the ND group (4.9 ± 0.4) (Table 1). Mean plasma protein CMC (μg/l) and CEC (μg/l) were also significantly higher in DM (55.73 ± 29.43 and 521.47 ± 239.13, respectively) compared to the ND group (24.25 ± 10.26 and 262.85 ± 132.02, respectively) (Fig. 4). In a correlation analysis, CMC correlated strongly with CEC (r = 0.903, P < 0.001), but HbA1C did not correlate well with either of the AGEs. Correlation analysis of data from DM group showed that both CMC and CEC were positively correlated with ACR (r = 0.442, P < 0.01 and r = 0.363, P < 0.01, respectively), as was HbA1C (r = 0.398, P < 0.01) (Table 2). In a multivariate linear regression analysis evaluating HbA1C, serum creatinine, and CMC or CEC as predictors of ACR, a combination of HbA1C and CMC (R 2 = 0.379, P < 0.001), or HbA1C and CEC (R 2 = 0.386, P < 0.001) were better predictors than any one of these variables alone (Table 2).

Table 1 Characteristics and biochemical parameters of study participants
Full size table
Fig. 4
figure 4

Plasma protein cysteine advanced glycation end products (AGEs), carboxymethyl cysteine (CMC), and carboxyethyl cysteine (CEC) of patients with diabetes (n = 42) and controls without diabetes (n = 18). Values are means ± SD. * indicates values are significantly different from the other group (P < 0.05)

Full size image
Table 2 Correlation of variables with ACR
Full size table

Discussion

We have developed a novel LC-MS/MS method to measure the cysteine AGEs, CMC and CEC, in plasma protein. To the best of our knowledge, this is the first report showing that these cysteine AGEs are elevated in the plasma of DM patients as compared to ND controls. Furthermore, we show that plasma protein CMC or CEC alone or in combination with HbA1C are good predictors of ACR, an indicator of nephropathy in patients with diabetes.

There is strong evidence that AGEs are elevated in diabetes and may be involved either directly, or indirectly via receptors, in the vascular complications of this condition [3–10]. The specific AGEs, CML, CEL and pentosidine, arise from modifications to lysine and/or arginine groups. These AGEs accumulate in serum, urine, skin collagen and kidney, and are associated with nephropathy in diabetes [3, 6–9]. They have also been shown to stimulate changes, which may contribute to diabetic vascular disease including increased oxidative stress [8, 34], proinflammatory activity [35–37], and endothelial dysfunction [38]. In addition to lysine and arginine AGEs, cysteine AGEs may also play a significant role in the vascular complications of diabetes. Cysteine residues are more powerful nucleophiles than lysine or arginine residues [19], which may make them more susceptible to AGE modification, and are found in many circulating and tissue proteins [20–26]. As glyoxal and methylglyoxal are elevated in individuals with diabetes [13], we anticipated that cysteine AGEs derived from these aldehydes would be high. Indeed we showed that plasma protein levels of CMC and CEC were elevated in the DM group as compared to ND controls.

Cysteine AGEs may exert detrimental effects on proteins in two ways. Firstly, circulating plasma protein CEC and CMC may act as ligands to RAGEs. These receptors are found on the cell surface of various cells such as endothelial, mesangial and vascular smooth muscle cells, neutrophils and podocytes [34–37, 39]. Stimulation of RAGEs can affect changes in intracellular signaling, protein expression, and cell growth [2]. Expression of RAGEs is increased in kidney of individuals with diabetes [6]. Secondly, like plasma proteins, other endogenous cysteine-containing proteins may be susceptible to CMC and CEC formation by glyoxal and methylglyoxal. This may influence the structure and function of various intra- and extracellular proteins that play a role in the progression of diabetic nephropathy. A role for AGEs and RAGEs in the etiology of kidney disease is supported by studies showing that treatment with RAGE antibodies, soluble RAGEs and AGE breakers attenuate adverse kidney changes in diabetes [2].

In the present study, the level of plasma protein CMC and CEC correlated reasonably well with ACR, an indicator of nephropathy, in patients with diabetes. Cysteine AGE-induced changes may contribute to inflammation, mesangial matrix expansion, thickening of glomerular basement membrane (GBM), albuminuria, and late-stage loss of filtering capacity characteristic of nephropathy. The GBM forms the barrier that prevents excretion of albumin and other macromolecules. In glomerular mesangial and endothelial cells, glycated albumin stimulates protein kinase C activity, increases growth factor bioactivity, and induces gene overexpression and production of matrix proteins [40]. In glomerular epithelial cells AGEs also decrease production of perlican core protein, a major heparin sulphate proteoglycan [41], which may result in a change in the anionic charge of GBM, leading to increased permeability and loss of albumin across this barrier. GBM contains cysteine residues, and in diabetes this membrane demonstrates a decrease in cysteine content [42]. This may be due to formation of cysteine AGEs with a decrease in this amino acid. AGE–RAGE interaction in podocytes increases production of growth and adhesion factors resulting in cell proliferation, matrix protein expression, and infiltration by inflammatory cells [43]. AGE-albumin shows a decreased binding and uptake by proximal tubule cells [44].

AGEs may also affect other cysteine-containing proteins that are involved in kidney function such as NOS. This enzyme catalyzes the formation of nitric oxide (NO) from arginine. NO is a vasodilator with anti-inflammatory properties and helps regulate sodium excretion in kidney [45, 46]. AGEs have been shown to reduce NOS activity in proximal tubular epithelial cells [47], and to decrease gene expression of this enzyme [38]. Decreased NOS in kidney vasculature reduces blood flow and sodium excretion, and increases blood pressure [46]. Disruption in NO availability also initiates a set of inflammatory/atherosclerotic responses [45] that may contribute to nephropathy. It has also been suggested that NO preserves the integrity of the glomerular permeability barrier to protein by antagonizing the effects of superoxide [48]. Thus a decrease in NO production would result in increased permeability and loss of albumin into the urine.

Oxidative stress is elevated in diabetes and may contribute to its vascular complications including nephropathy [8, 49]. The activity of antioxidant enzymes, glutathione peroxidase, and reductase are dependent on selenocysteine and cysteine residues, respectively [22], making them susceptible to AGE modification. Methylglyoxal has been shown to inhibit the activity of both these enzymes [21, 27]. AGEs have been shown to induce oxidative stress in mesangial cells [34] and macrophages [50]. AGE–RAGE interaction also increases oxidative stress via NADPH-oxidase in endothelial cells [51]. Thus, cysteine AGE-induced oxidative stress may be responsible in part for tissue damage associated with nephropathy.

Plasma HbA1C is a measure of structural modification to haemoglobin as a result of hyperglycemia, and reflects short-term glycemic control. In the present study, although HbA1C correlated with ACR, it did not correlate well with either plasma protein CMC or CEC. It is likely that the increases in methylglyoxal and glyoxal seen in patients with diabetes arise not primarily from hyperglycemia per se, but rather from altered intracellular glucose and lipid metabolism. If, as we suggest, CMC and CEC are involved in the etiology of diabetic nephropathy, levels of these AGEs in plasma protein may be a more direct representation of structural and functional changes to kidney tissue. Plasma protein CEC or CMC in combination with HbA1C were better predictors of ACR than either of these variables individually, indicating that glucose, as well as glyoxal and methylglyoxal, may be involved in diabetic nephropathy. This suggests the importance of dietary and pharmaceutical therapies, which not only normalize glucose levels, but control levels of AGEs, in individuals with diabetes.

We acknowledge certain limitations for this study. As routine blood sampling was used to collect plasma for measurement of CMC and CEC in our control group without diabetes, serum and urine samples were not available for this group. As our measurements for HbA1C in these patients were normal, then it is reasonable to expect that serum and urine creatinine, and microalbuminuria measurement would be within normal range in these patients. We measured only two cysteine AGEs. As we continue to develop our methodology to identify and measure other AGEs that may be involved in nephropathy, we may be able to better predict nephropathy in the patients with diabetes.

Conclusion

We have developed a novel LC-MS/MS method to measure plasma protein CMC and CEC, AGEs formed from the aldehydes glyoxal and methylglyoxal, respectively. We are the first to show that the levels of these AGEs in plasma proteins are elevated in a group of patients with diabetes as compared to a group of controls without diabetes. Furthermore, we show that plasma protein CMC and CEC correlated well with ACR, an indicator of nephropathy in patients with diabetes, suggesting a possible etiological role for these AGEs in this vascular complication. CMC or CEC in combination with HbA1C are better predictors of nephropathy. This may indicate that glucose, glyoxal, and methylglyoxal may all be involved in the etiology of diabetic nephropathy.