Development and Application of a Stable Isotope Dilution Analysis

Annie R. Knight , Emma Louise Taylor , Jeannette Oettrich , Tatjana Ruskovska , Ana Cipak ... Patrick Giraudeau , Illa Tea , Gérald S. Remaud , S...
0 downloads 0 Views 702KB Size
Article pubs.acs.org/ac

Development and Application of a Stable Isotope Dilution Analysis for the Quantitation of Advanced Glycation End Products of Creatinine in Biofluids of Type 2 Diabetic Patients and Healthy Volunteers Christof Kunert,† Thomas Skurk,‡,§ Oliver Frank,† Roman Lang,† Hans Hauner,‡,§ and Thomas Hofmann*,†,‡ †

Chair of Food Chemistry and Molecular Sensory Science, Technische Universität München, Lise-Meitner-Strasse 34, D-85354 Freising-Weihenstephan, Germany ‡ Research Center for Nutrition and Food Sciences (ZIEL), Clinical Nutritional Medicine, Technische Universität München, Gregor-Mendel-Strasse 2, D-85350 Freising, Germany § Klinikum Rechts der Isar, Technische Universität München, Ismaninger Straße 22, D-81675 München, Germany S Supporting Information *

ABSTRACT: N-(1-Methyl-4-oxoimidazolidin-2-ylidene) α-amino acids were recently identified in roasted meat as so far unknown advanced glycation end products (AGEs) of creatinine. For the first time, this paper reports on the preparation of 13C-labeled twin molecules of six N-(1-methyl-4-oxoimidazolidin-2-ylidene) αamino acids and the development of a stable isotope dilution analysis (SIDA) for their simultaneous quantitation in meat, plasma, and urine samples by means of HPLC-MS/MS. Method validation demonstrated good precision (99% by means of HPLC-ELSD. To account for water residues, the accurate concentration of solutions of compounds 1−6 and their 13C-isotopologues was determined by means of quantitative nuclear magnetic resonance (qNMR) spectroscopy using the proton signal at 7.62 ppm recorded for the calibrating reference benzoic acid and the resonance signal of H−C(3) of the creatinine moiety of 13 Cn-labeled and natural 13C-abundant compounds 1−6, respectively. The qNMR method revealed excellent linearity and precision for the calibrant benzoic acid (y = 0.9966x + 0.0077, 0.9999 < R2 < 1.0000). Development of a Stable Isotope Dilution Analysis (SIDA). For tuning the mass spectrometer, solutions of the reference compounds 1−6 and their 13C-labeled isotopologues,

increased to 100% within 1 min, then held for 3 min, decreased to starting conditions within 1 min, and held for 7 min prior to the next injection.



RESULTS AND DISCUSSION

Recently, the N-(1-methyl-4-oxoimidazolidin-2-ylidene) αamino acids 1−6, Figure 1, were isolated from food-related Maillard reaction systems of creatinine and reducing carbohydrates. As compounds 1a/b and 6a/b were identified as taste modulators in thermally processed beef meat,25,26 the question arose as to whether these N-(1-methyl-4-oxoimidazolidin-2ylidene) α-amino acids are absorbed after ingestion of thermally processed meat and/or are also formed in vivo under physiological conditions by nonenzymatic glycation of creatinine. Therefore, a stable isotope dilution analysis (SIDA) should be developed for the quantitation of 1−6 in meat samples and biofluids (urine, plasma). To meet this demand, stable isotope labeled twin molecules of 1−6 were synthesized as internal standards. Synthesis of 13C-Labeled Isotopologues of Compounds 1−6. Binary mixtures of creatinine and 13C5-ribose 2965

dx.doi.org/10.1021/ac303684v | Anal. Chem. 2013, 85, 2961−2969

Analytical Chemistry

Article

Figure 3. Generation of creatinine glycation products 1−6 upon incubation (37 °C, pH 7.4) of creatinine with 2-oxopropanal (), ribose (- - -), and glucose (− − −), respectively; error bars express the standard error of the mean (SEM).

evaluation of accuracy and precision, an aqueous raw meat extract serving as plasma-like matrix was spiked with the analytes 1−6 in five concentrations between 0.10 and 2.00 μg/ L. Precision, expressed as relative standard deviation of replicate analysis (n = 4), and accuracy in accordance with nominal and measured concentration is given in Table 2. Precision adopted values between 2% and 14% for all analytes, and accuracy was between 97% and 118% across the studied concentration range, indicating the developed SIDA-LC-MS/MS method as a reliable tool enabling an accurate quantitative determination of 1−6. Since analysis of urine involved only internal standard addition and dilution prior to analysis, validation experiments were focused on plasma analysis only. Quantitation of 1−6 in Physiological Model Incubations. To check whether compounds 1−6 are formed under physiological conditions (37 °C, pH 7.4), creatinine was incubated with glucose, ribose, or 2-oxopropanal, a well-known carbohydrate fragmentation product,12 and after 0, 3, 7, 14, 21, and 28 days, samples were withdrawn and analyzed for the formation of N-(1-methyl-4-oxoimidazolidin-2-ylidene) α-

respectively, were individually infused into the ion source of the MS system to optimize the ion intensities of the pseudomolecular ion [M + H]+, and the fragments were generated by collision induced dissociation by means of software-assisted ramping of the ion source and ion path potentials. The most intensive mass transition was used for quantitation, and the second transition was selected for unequivocal identification of the target analytes (Table 1). Sufficient separation of 1−6 was achieved on a PFP-column using acidified water and acetonitrile as eluents (Figure 2). Calibration curves were evaluated in the range of 0.10−100 μg/L with a fixed concentration of 20 μg/L of the respective internal standards. The linear range was found in a concentrations range from 0.10 to 100 μg/L for compounds 1−6 with R2 values between 0.993 and 0.999 for all the compounds. Precision of the backcalculated standards was 2.6−12.9%; accuracy was 93.2− 117.4%. The lower limit of quantitation (LLOQ), calculated from the injection volume and the lowest standard that fell in the acceptance criteria (precision 200 μM already after 3 days of incubation (1 in Figure 3). In comparison, generation of 1 from the reaction of creatinine with glucose and ribose, respectively, proceeded much slower; e.g., levels of about 200 μM were found after 28 days. The higher homologue N-(1-methyl-4-oxoimidazolidin-2-ylidene) aminobutanoic acid (2) was only formed in comparatively low amounts reaching a maximum concentration of about 7 μM when creatinine was reacted with 2-oxopropanal (2 in Figure 3). N-(1-Methyl-4-oxoimidazolidin-2-ylidene) aminohydroxybutanonic acid (3) was found as the quantitatively predominant reaction product with concentrations up to 1000 μM in creatinine/2-oxopropanal incubations and levels of about 800 and 700 μM when creatinine was reacted with glucose and ribose, respectively (3 in Figure 3). In comparison, the higher homologues 4−6 (4, 5, and 6 in Figure 3) were found only in model incubations of creatinine and glucose or ribose with maximum concentrations of about 110 μM for 5 in the creatinine/ribose model (5 in Figure 3). As these investigations revealed the formation of N-(1-methyl-4-oxoimidazolidin-2-

ylidene) amino acids (1−6) under pseudophysiological conditions, the question arose as to whether these compounds are formed in vivo as well. Therefore, a clinical human intervention study was conducted with type 2 diabetes patients and healthy volunteers in order to investigate whether the bioappearance of compounds 1−6 mirrors the dietary uptake of N-(1-methyl-4-oxoimidazolidin-2-ylidene)-α-amino acids and whether the endogenous formation of 1−6 in vivo is occurring to a significant extend. Human Intervention Study. To get first insights into the kinetics concerning dietary uptake and/or in vivo formation of N-(1-methyl-4-oxoimidazolidin-2-ylidene) α-amino acids, a human intervention study was designed with 7 volunteers with diabetes mellitus (DM) type 2 taking oral hypoglycemic medication or insulin and 10 healthy volunteers. Within a period of 7 days prior to the intervention study, all participants were asked to abstain from any meat and meat containing products, respectively. After this wash-out phase, the volunteers were fasted for 12 h, and blood and urine samples were collected prior to the intervention. Thereafter, participants consumed two beef burgers comprising a total of ∼342 g of meat on average within 30 min. SIDA analysis of N-(1-methyl4-oxoimidazolidin-2-ylidene) α-amino acids in the ingested beef 2967

dx.doi.org/10.1021/ac303684v | Anal. Chem. 2013, 85, 2961−2969

Analytical Chemistry

Article

meat samples revealed concentrations of 129.79 ± 1.20 mg/kg for 1, 0.96 ± 0.11 mg/kg for 2, and 10.08 ± 0.26 mg/kg for 3, whereas compounds 4−6 could not be detected. The absolute ingested amounts of compounds 1−3, therefore, were 44.5 mg (1), 0.33 mg (2), and 3.45 mg (3). Following the food intake, additional blood and urine samples were taken after 1, 2, 4, 6, 9, and 24 h, respectively. Aliquots of the urine excreted (DM patients: mean 2108 ± 623 mL; healthy control: mean 2829 ± 512 mL) and blood samples, after centrifugation at 3000g (10 min, 4 °C), were used for HPLC-MS/MS analysis. Application of the SIDA on plasma samples collected prior to the intervention revealed N-(1-methyl-4-oxoimidazolidin-2ylidene) α-aminopropionic acid (1) as the only abundant metabolite at fasting levels, whereas the derivatives 2−6 could not be detected (Table 3). This is the first report on the presence of 1 in the bloodstream. The basal, fasting concentration of 1 in the plasma of diabetes patients was significantly higher (2.05 ± 0.41 μg/L; p < 0.01, two-tailed ttest) compared to healthy volunteers (0.63 ± 0.12 μg/L). Analysis of the morning urine samples showed that 1 was the dominating analyte but small amounts of derivates 2 and 3 were found as well. Comparing the absolute excreted amounts of 1 in the morning urine, no significant differences were detected between healthy and DM volunteers (p = 0.70, two-tailed ttest), indicating reduced renal filtration of the compound in DM patients. However, as shown in Figure 4, the meat ingestion induced a remarkable increase in urinary concentrations of compound 1 in both study groups. Levels in DM patients (filled circles) increased about 18-fold from ∼3.17 to ∼57 μg/L (Figure 4A), corresponding to absolute excreted amounts ranging from 0.36 μg in the morning urine to 14 μg after 8 h (Figure 4D). In the healthy study participants, the excreted amount of compound 1 was significantly higher even 24 h after meat ingestion compared to the basal level with 0.42 ± 0.13 μg (time 0) compared to 1.56 ± 0.27 μg (24 h, p = 0.0016). In contrast, no significant differences were detected in the DM patients (0.36 ± 0.13 μg at t = 0 compared to 2.01 ± 0.68 μg at t = 24 h, p = 0.0719). In comparison, the concentrations found in the urine of the healthy volunteers (open boxes) was raised only 9-fold from 1.82 μg/L (0.42 μg absolute amount excreted) in morning urine to a maximum of 17 μg/L, accounting for an absolute excreted amount of 4 μg (Figure 4A,D). Similar time/concentration profiles were observed for compounds 2 (Figure 4B,E) and 3 (Figure 4C,F), but compared to 1, urinary concentrations of N-(1-methyl-4oxoimidazolidin-2-ylidene) aminobutanoic acid (2) under fasting conditions were lower with levels of 0.03 (±0.01) and 0.06 μg/L (±0.02) in healthy volunteers and diabetes patients, respectively (Figure 4B,E, Table 3). Whereas urine levels of 2 in DM patients went through a maximum of about 0.69 μg/L (±0.17) at 2 h after meat intake, the concentrations in healthy subjects reached a maximum of only 0.29 μg/L (±0.07); thereafter, the levels of 2 declined again to reach starting levels after about 24 h (Figure 4B,E). Furthermore, also urinary levels of N-(1-methyl-4-oxoimidazolidin-2-ylidene) aminohydroxybutanoic acid (3) increased from 0.45 μg/L (±0.14) to 2.59 μg/L (±0.62) in diabetes patients, whereas in healthy controls the baseline level of 0.30 μg/L (±0.07) showed only a marginal increase to a concentration of 0.75 μg/L (±0.30) about 9 h after meat ingestion (Figure 4C,F). Although only compound 1 could be unequivocally detected in plasma, the difference observed in the urinary excretion

pattern between healthy and diabetes volunteers was reflected in the plasma. While compound 1 increased from 2.05 to 6.91 μg/L within 4 h after meat ingestion in diabetes patients (filled circles), the plasma peak was 3.63 μg/L in healthy volunteers (open boxes). Diabetic patients already presented elevated basal levels compared to nondiseased (2.05 ± 0.41 vs 0.63 ± 0.12 μg/L, respectively; p < 0.01). After 24 h, no significant difference between the plasma concentrations measured before and after meat ingestion were found (two-tailed t-test, p = 0.305 (DM) and p = 0.302 (healthy), Figure 5A). Although our

Figure 5. Concentration/time profiles of (A) N-(1-methyl-4oxoimidazolidin-2-ylidene) α-aminopropionic acid (1) and (B) creatinine in plasma of diabetes type 2 (DM) patients (●) and healthy controls (□) collected after meat ingestion (t0) over 24 h; error bars give the standard error of the mean (SEM).

approach could not discriminate between elevated resorption and endogenous production, the higher concentration of another AGE entity in the plasma was very recently shown to be predictive for the impairment of vascular function in diabetes.30 It is obvious that the intervention only slightly affected plasma creatinine levels in healthy volunteers while in diabetes patients an initial increase from 4420 to 5200 μg/L was observed. Interestingly, the plasma creatinine concentration was significantly lower in DM volunteers compared to the healthy controls (two tailed student’s t-test, P = 0.002) (Figure 5B). This is in line with recently published reports showing that low serum creatinine is associated with an increased diabetes type 2 risk. The reason for this observation is vastly unknown but reduced lean mass or renal hyperfiltration were hypothesized as pathogenic factors.2,31 Thus, it is highly interesting to notice that levels of glycated creatinine in DM patients by far exceeds the increase observed in healthy subjects after a meal containing beef. 2968

dx.doi.org/10.1021/ac303684v | Anal. Chem. 2013, 85, 2961−2969

Analytical Chemistry



Article

(11) Yeh, C. H.; Sturgis, L.; Haidacher, J.; Zhang, X. N.; Sherwood, S. J.; Blercke, R. J.; Juhasz, O.; Crow, M. T.; Tilton, R. G.; Denner, L. Diabetes 2001, 50, 1495−1504. (12) Severin, T.; Hiebl, J.; Popp-Ginsbach, H. Z. Lebensm. Unters. Forsch. 1984, 178, 284−287. (13) Wells-Knecht, K. J.; Zyzak, D. V.; Litchfield, J. E.; Thorpe, S. R.; Baynes, J. W. Biochemistry 1995, 34, 3702−3709. (14) Reddy, S.; Bichler, J.; Wells-Knecht, K. J.; Thorpe, S.; Baynes, J. Biochemistry 1995, 34, 10872−10878. (15) Teerlink, T.; Barto, R.; Brink, H.; Schalkwijk, C. Clin. Chem. 2004, 50, 1220−1228. (16) Yagmur, E.; Tacke, F.; Weiss, C.; Lahme, B.; Manns, M. P.; Kiefer, P.; Trautwein, C.; Gressner, A.M.. Clin. Biochem. 2006, 39, 39− 45. (17) Jono, T.; Nagai, R.; Lin, X.; Ahmed, N.; Thornalley, P. J.; Takeya, M.; Horiuchi, S. J. Biochem. 2004, 136, 351−358. (18) Caughlan, M.; Forbes, J. Am. J. Nephrol. 2011, 34, 9−17. (19) Sternberg, Z.; Hennies, C.; Sternberg, D.; Wang, P.; Kinkel, P.; Hojnacki, D.; Weinstock-Guttmann, B.; Munschauer, F. J. Neuroinflammation 2010, 7:72, 1−8. (20) Lederer, M. O.; Gerum, F.; Severin, T. Bioorg. Med. Chem. 1998, 6 (7), 993−1002. (21) Anacardio, R.; Cantalini, M. G.; Angelis, F.; Gentile, M. J. Mass Spectrom. 1997, 32, 388−394. (22) Mostafa, A.; Randell, E.; Vasdev, S.; Gill, V. D.; Han, Y.; Gadag, V.; Raouf, A. A.; El Said, H. Mol. Cell. Biochem. 2007, 302, 35−42. (23) Hielmesaeth, J.; Roislien, J.; Nordstrand, N.; Hofso, D.; Hager, H.; Hartmann, A. BMC Endocr. Disord. 2010, 10:6, 1−6. (24) Aso, Y.; Inukai, T.; Tayama, K.; Takemura, Y.. Acta Diabetol. 2000, 37, 87−92. (25) Kunert, C.; Walker, A.; Hofmann, T. J. Agric. Food Chem. 2011, 59, 8366−8374. (26) Sonntag, T.; Kunert, C.; Dunkel, A.; Hofmann, T. J. Agric. Food Chem. 2010, 58, 6341−6350. (27) Wider, G.; Dreier, L. J. Am. Chem. Soc. 2006, 128, 2571−2576. (28) Korn, M.; Frank, O.; Hofmann, T.; Rychlik, M. Anal. Biochem. 2011, 419, 88−94. (29) Lang, R.; Wahl, A.; Stark, T.; Hofmann, T. Mol. Nutr. Food Res. 2011, 55, 1613−1623. (30) Stirban, A.; Kotsi, P.; Franke, K.; Strijowski, U.; Cai, W.; Götting, C.; Tschoepe, D. Diabetes Care, published ahead of print on 13, December 2012; DOI: 10.2337/dc12-1489. (31) Andreev, E.; Koopman, M.; Arisz, L. J. Intern. Med. 1999, 246, 247−252.

CONCLUSION In conclusion, a robust and sensitive SIDA-HPLC-MS/MS method was developed for the quantitation of N-(1-methyl-4oxoimidazolidin-2-ylidene) α-amino acids (1−6) in urine and plasma samples. Application of this method to a human intervention study clearly demonstrated for the first time significant differences in the response to a meat containing meal between healthy subjects and patients with type 2 diabetes mellitus. Urine and plasma samples collected from type 2 diabetes patients showed elevated levels of N-(1-methyl-4oxoimidazolidin-2-ylidene) α-amino acids when compared to healthy controls. After a 7 day meat-free diet and a 12 h fasting period (time point t0), already baseline levels of 1 in DM patients were elevated and a significantly higher concentration after meat ingestion compared to healthy controls was observed in plasma and urine. Furthermore, excretion of 1−3 in urine was significantly pronounced in the diseased population. The data suggest further speculation that nonenzymatically glycosylated creatinine could be involved in the postprandial impairment of macrovascular function. Using the SIDA-LCMS/MS method developed here, future studies need to unravel whether N-(1-methyl-4-oxoimidazolidin-2-ylidene) α-amino acids such as, e.g., 1, could serve as new biomarkers assessing the vascular risk profile after meat ingestion. Moreover, clarity is needed on the mechanisms underlying the observed effects: increased dietary uptake, reduced excretion, and/or enhanced in vivo formation of N-(1-methyl-4-oxoimidazolidin-2-ylidene) α-amino acids.



ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: +49-8161/71-2902. Fax: +49-8161/71-2949. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Brownlee, M.; Cerami, A.; Vlassara, H. N. Engl. J. Med. 1988, 318, 1315−1320. (2) Harita, N.; Hayashi, T.; Sato, K. K.; Nakamura, Y.; Yonedo, T.; Endo, G.; Kambe, H. Diabetes Care 2009, 32, 424−426. (3) Vlassara, H.; Palace, M. R. J. Intern. Med. 2002, 251, 87−101. (4) Miyata, T.; Hori, O.; Zhan, J.; Yan, S. D.; Ferran, L.; Iida, Y.; Schmidt, A. M. J. Clin. Invest. 2001, 98, 1088−1094. (5) Thornalley, P. J. Endocrinol. Metab. 1996, 3, 149−166. (6) Neeper, M.; Schmidt, A. M.; Brett, J.; Yan, S. D.; Wang, F.; Pan, Y. C.; Elliston, K.; Stern, D.; Shaw, A. J. Biol. Chem. 1992, 267, 14998− 15004. (7) Schmidt, A. M.; Vianna, M.; Gerlach, M.; Brett, J.; Ryan, J.; Kao, J.; Esposito, C.; Hegarty, H.; Hurley, W.; Clauss, M. J. Biol. Chem. 1992, 267, 14987−14997. (8) Yan, S. D.; Schmidt, A. M.; Anderson, G. M.; Zhang, J.; Brett, J.; Zou, Y. S.; Pinsky, D.; Stern, D. J. Biol. Chem. 1994, 269, 9889−9897. (9) Lander, H. M.; Tauras, J. M.; Ogiste, J. S.; Hori, O.; Moss, R. A.; Schmidt, A. M. J. Biol. Chem. 1997, 272, 17810−17814. (10) Taguchi, A.; Blood, D. C.; del Toro, G.; Canet, A.; Lee, D. C.; Qu, W.; Tanji, N.; Lu, Y.; Lalla, E.; Fu, C.; Hofmann, M. A.; Kislinger, T.; Ingram, M.; Lu, A.; Tanaka, H.; Hori, O.; Ogawa, S.; Stern, D. M.; Schmidt, A. M. Nature 2000, 405, 354−360. 2969

dx.doi.org/10.1021/ac303684v | Anal. Chem. 2013, 85, 2961−2969