Bioorganic & Medicinal Chemistry Letters xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Bioorganic & Medicinal Chemistry Letters journal homepage: www.elsevier.com/locate/bmcl

Generation and characterization of antibodies against arginine-derived advanced glycation endproducts Tina Wang, Matthew D. Streeter, David A. Spiegel ⇑ Department of Chemistry, Yale University, 225 Prospect Street, New Haven, CT 06511, United States

a r t i c l e

i n f o

Article history: Received 1 May 2015 Accepted 2 June 2015 Available online xxxx Keyword: Advanced glycation endproducts

a b s t r a c t Although antibodies reagents have been widely employed for studying advanced glycation end-products (AGEs), these materials have been produced using complex mixtures of immunogens. Consequently, their epitope specificity remains unknown. Here we have generated the first antibodies capable of recognizing each of the three isomers of the methylglyoxal hydroimidazolones (MG-Hs) by using chemical synthesis to create homogenous immunogens. Furthermore, we have thoroughly characterized the epitope specificity of both our antibodies and that of two existing monoclonals by implementing a direct ELISA protocol employing synthetic MG-H antigens. Finally, we employed the reported anti-MG-H antibodies to the detection of MG-Hs in cellular systems using immunofluorescence microscopy. These studies have demonstrated that anti-MG-H1 and anti-MG-H3 staining is concentrated within the nucleus, while anti-MG-H2 affords only minimal signal. These observations are consistent with reported formation preferences for MG-Hs, and may suggest novel nuclear targets for non-enzymatic posttranslational modification. The antibody reagents reported herein, as well as the strategy employed for their creation, are likely to prove useful for the immunochemical study of AGEs in biological systems. Ó 2015 Elsevier Ltd. All rights reserved.

Introduction Advanced glycation endproducts (AGEs) are posttranslational modifications formed by the non-enzymatic reaction of protein side-chains with sugars and sugar degradation products.1,2 AGE levels have been shown to be elevated during the aging process,3 and in various disease states, such as diabetes,4 cancer,5,6 and cardiovascular disease.7 Indeed, several AGEs have been shown to serve as useful diagnostic markers of disease.8–10 In addition, AGEs are thought to contribute directly to disease pathophysiology by causing protein damage and dysfunction. AGEs have also been shown to induce a number of deleterious effects in cellular systems, such as induction of oxidative stress and pro-inflammatory signaling.1 Thus, there is considerable interest in the role(s) of AGEs in human health and disease. The methylglyoxal hydroimidazolones (MG-Hs) comprise the most prevalent arginine-derived AGE.11 They are formed as a mixture of three isomers by protein glycation with methylglyoxal (MGO), a byproduct of glycolysis (Fig. 1A). MG-Hs are estimated to modify 1–2% of all arginine residues found in lens proteins of elderly human subjects.12 Levels of this AGE have also been found to be elevated in patients with cardiovascular disease,13 ⇑ Corresponding author. Tel.: +1 (203) 432 8697; fax: +1 (203) 432 6144. E-mail address: [email protected] (D.A. Spiegel).

Alzheimer’s disease,14 and diabetes mellitus (DM).15,16 Moreover, with respect to DM, increased MG-H levels are strongly correlated with the onset of complications such as retinopathy,17 highlighting the potential utility of these modifications as biomarkers. Many techniques are available for the visualization of the MGHs and other AGEs, including UV–Vis and fluorescence imaging,18,19 exhaustive hydrolysis followed by HPLC,20,21 and mass spectrometry.22,23 However, these methods suffer from various limitations. For example, UV–Vis and fluorescence detection are usually limited to bulk analysis, and require AGEs that possess suitable chromophores. Chromatography and mass spectrometry are useful in their ability to identify specific AGEs in biological samples, but require specialized instrumentation, and are expensive with respect to both cost and time. Antibodies have been widely utilized in the study of AGEs, and have enabled histological and immunochemical experiments, providing an attractive alternative to chromatography/mass spectrometry for detecting AGEs in complex biological samples. Despite their widespread usage, however, in general, the epitopic specificity of these reagents is not known. This problem arises for two main reasons. First, immunizing antigens are often produced by reacting a carrier protein (albumin or KLH) with glycating agents, leading to a complex mixture with unknown AGE composition. Therefore, although the resulting MGO-protein mixtures are often assumed to exclusively contain the various MG-Hs,

http://dx.doi.org/10.1016/j.bmcl.2015.06.013 0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Wang, T.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.06.013

2

T. Wang et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx

A

Me O

HN

O

Me

O N

N H

protein NH

N

protein NH

protein

protein MG-H1

MG-H2

O N

protein NH

N H

protein

Me

NH

N

H 2N

NH 2 Me

O

O

protein

O N

H 2N

O

N H

O

protein

MG-H3

CEA

B

X

H

H N

1, X = Arg H 2N

NH 2 NH

Arg

O N H

O

H N

N H

O

2, X = MG-H1 H N

4 O

3, X = MG-H2 O

Me

N

HN

H N

O

O

MG-H1 Me

N

N

MG-H2 NH 2

CO 2H

SH 2 4, X = MG-H3 Me

O N

N

MG-H3 NH 2

Figure 1. (A) Structures of the methylglyoxal hydroimidazolone (MG-H) isomers and MG-H3 hydrolysis product carboxyethylarginine (CEA). (B) MG-H and argininemodified peptide constructs used for immunization and antibody purification.

numerous other AGE adducts have also been isolated from these preparations.11 Second, because chemically-homogeneous synthetic AGE standards have not historically been available for most AGEs, the exact structure of recognition epitopes cannot be rigorously determined. Instead, these antibodies are generally characterized by showing preferential binding to AGE-modified proteins over non-modified counterparts. Such strategies are inherently imprecise because these AGE-proteins contain a heterogeneous array of AGEs. Indeed, while antibodies have been reported to recognize CML,24–26 CEL,27,28 pentosidine,29 and argpyrimidine,30 for many anti-AGE antibodies, the actual epitope(s) being recognized is not clear.31 With respect to the MG-Hs, two mouse monoclonal antibodies, clones 3D11 and 1H7G5 (also referred to as IG7),32,33 have been reported. Both 3D11 and 1H7G5 were generated using immunogen created by incubation of carrier protein with MGO,32,34 and these reagents have been used extensively in the literature to study MGO-protein modifications. While 1H7G5 is thought to recognize MG-H1, and to a lesser extent argpyrimidine,34 the epitopic specificity of 3D11 has not been described in the literature. Wellcharacterized antibody reagents capable of recognizing the three MG-H isomers independently would therefore be highly useful, both as diagnostics and as research tools. Herein we report the generation and characterization of the first selective antibodies against each of the three MG-H isomers. These efforts were enabled by our laboratory’s recent development of efficient syntheses for each of these isomers as both amino acids and peptide conjugates.35 Thus, we have constructed chemically homogeneous MG-H preparations, and used these to immunize rabbits for production of selective polyclonal sera. We have also thoroughly characterized the epitopic specificity of both our antibodies, and the two existing anti-MG-H monoclonals, through

the use of a direct ELISA employing synthetic antigens. Interestingly, while both anti-MG-H1 and anti-MG-H2 sera proved highly selective for the target antigens, anti-MG-H3 was found to cross-react with carboxyethyl arginine (CEA). Monoclonal antibody 1H7G5 exhibited a similar pattern of cross-reactivity between these two antigens while 3D11 cross-reacted with MG-H1, MG-H3, and CEA. Finally, immunofluorescence microscopy experiments indicated MGO-dependent increases in staining with antiMG-H1 and anti-MG-H3, but not anti-MG-H2 sera. The staining pattern observed for these antibodies was predominantly nuclear, although faint cytosolic staining was also visualized. This Letter is the first to demonstrate the generation of selective antibodies against individual MG-Hs. These new reagents have the potential to serve as useful tools for both disease diagnosis and fundamental research. The MG-H class of AGEs is composed of three isomers, as illustrated in Figure 1A. Our goal was to generate antibodies capable of recognizing each MG-H isomer while exhibiting minimal crossreactivity with the other two. To accomplish this, we designed the immunogens depicted in Figure 1B. Thus, for each isomer, synthetic peptide immunogens were constructed to contain the MG-H modification flanked by glycine and spaced out from the terminal cysteine residue (the site of immunogen protein conjugation) with a flexible polyethylene glycol (PEG) linker (2–4, Fig. 1B). We employed amino acids with minimal side-chain functionality in order to obtain antibodies capable of recognizing their epitope in a sequence-independent manner. These immunogens were produced using solid phase peptide synthesis (SPPS), purified, and conjugated to keyhole limpet hemocyanin (KLH) for immunization. Antibodies were then isolated through a series of negative and positive affinity purifications. For example, to obtain MG-H3 specific antibodies, the antiserum was first depleted using 1 to remove

Please cite this article in press as: Wang, T.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.06.013

T. Wang et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx

any fraction that cross-reacted with either the immunogen scaffold or unmodified arginine residues. Then, antibodies recognizing the MG-H1 and MG-H2 isomers were removed by further negative affinity purification against 2 and 3. Finally, the remaining antibodies were enriched for MG-H3 binders by positive affinity purification with the original immunogen 4. We then tested the specificities of these polyclonal sera using an enzyme-linked immunosorbent assay (ELISA) protocol. For these assays, we synthesized biotinylated MG-H containing peptides with a parent sequence derived from human serum albumin (HSA). This sequence was chosen to be completely unrelated to that of the immunogens, thus allowing us to determine if the purified antibodies could recognize MG-H epitopes in different peptide contexts. These biotinylated peptides were bound directly to avidin-coated plates, and antibody binding was then evaluated using direct ELISA. For comparison, we also evaluated the epitope recognition of two existing monoclonal antibodies that have been generated using MGO-modified proteins (clones 3D11 and 1H7G5). In the direct ELISA format, our antibodies bound their respective MG-Hs with EC50s in the low nanomolar to high picomolar range, even though in each case, the carrier peptide sequence was completely different from that of the immunogen (Table 1, entries 1–3). Monoclonals 3D11 and 1H7G5 also bound MG-H epitopes with similar affinities (Table 1, entries 4 and 5). All antisera tested showed negligible binding to the unmodified control peptide, as well as to the avidin and BSA used for coating and blocking, respectively. The MG-H2 antiserum possessed the best specificity, showing little or no cross-reactivity with the other MG-H isomers or carboxyethylarginine (CEA), an intermediate that has been shown to arise in vitro by MG-H3 hydrolysis (Fig. 2B; Table 1, entry 2).21 MG-H1 antiserum exhibited some cross-reactivity with MG-H2 and CEA at high dilutions, but was approximately 5-fold more specific for MG-H1 at lower dilutions (Fig. 2A; Table 1, entry 1). Finally, while MG-H3 antiserum did not recognize other MG-H isomers, we observed that it strongly bound to CEA with an EC50 approximately 10-fold lower than that of MG-H3 (Fig. 2C; Table 1, entry 3). We believe this cross-reactivity arises from in vivo hydrolysis of the MG-H3 immunogen into CEA, a process that has been reported to have a half-life of less than a day under physiological conditions.21 It is unclear whether the antibodies consist of separate MG-H3 and CEA reactive pools, or if they are capable of binding both epitopes (or a mixture of both). Attempts to negatively purify CEA-reactive antibodies from this pool resulted in greatly reduced antibody yield, suggesting that a large fraction of these antibodies recognize CEA. With respect to the monoclonal antibodies tested, 3D11 recognized MG-H1 and MGH3 with near-equal affinity, and also strongly cross-reacted with CEA (Fig. 2D; Table 1, entry 4). Monoclonal 1H7G5, on the other hand, was found to recognize MG-H3 and CEA with near-equal affinity, but was insensitive to the other isomers (Fig. 2E; Table 1, entry 5). Table 1 EC50s of antibody binding to peptide sequence K(biotin)-LLV-X-YTKKV, X = R or MGH/CEA modification, measured by direct ELISA Entry

Antibody

1 2 3 4 5

Anti-(MG-H1) Anti-(MG-H2) Anti-(MG-H3) 3D11 1H7G5

EC50 ± SD (nM)a MG-H1

MG-H2

MG-H3

CEA

0.90 ± 0.32 n.d. n.d. 0.11 ± 0.04 n.d.

14.2 ± 4.2 0.80 ± 0.06 n.d. n.d. n.d.

n.d. b n.d. 3.1 ± 1.3 0.21 ± 0.11 1.6 ± 0.05

21.4 ± 2.3 n.d. 0.22 ± 0.05 0.98 ± 0.57 3.7 ± 1.2

a EC50s reported as mean ± error; error represents standard deviation of two independent experiments. b EC50s not determined due to low affinity (lack of plateau at highest concentrations of antibody tested).

3

Thus, we succeeded in generating polyclonal antisera that bound each of the MG-H isomers with high affinity and varying levels of specificity. Using our direct ELISA, we also were able to characterize the epitope specificity of two existing monoclonal antibodies that have been previously used to measure MGO modifications, and found that each recognized at least two different arginine-derived AGEs. Interestingly, of all the reagents tested, only the anti-MG-H1 antiserum produced exhibited specific recognition of MG-H1. This observation contrasts with previous reports claiming that monoclonal antibody 1H7G5 recognizes MG-H1.34 Having confirmed that our antibodies successfully recognize MG-H isomers in the context of synthetic peptides, we sought to test if these antibodies could recognize protein-bound MG-Hs. To this end, AGE-BSA,36 CML-BSA, minimally MGO-modified BSA and HSA,37 and their unmodified controls were coated on 96-well plates and MG-H antibody binding was evaluated using direct ELISA. In this assay, both MG-H1 and MG-H3 antibodies were found to bind MGO-modified albumin while exhibiting minimal binding to unmodified control albumin (Fig. 2F). However, MGH2 antibodies exhibited considerably less binding to MGO-modified protein, suggesting that MG-H2 is not a dominant epitope on MGO-modified protein. This is consistent with previous observations that this isomer is not formed in the reaction between arginine and MGO.11,20 Interestingly, MG-H1 is reported to be the predominant modification on ‘minimally modified’ MGO-HSA (>90% of total MGO-derived modifications detected).37 However, MG-H3 antibody, which exhibits virtually no cross-reactivity with MG-H1, also strongly binds this protein (Fig. 2F). This observation suggests that either MG-H3 or CEA (or both) also comprise a significant portion of the modifications on MGO-HSA, and on MGO-modified proteins in general. We speculate that MG-H3 modifications may have been overlooked in previous work because of their recently-demonstrated tendency to oxidize spontaneously, and decompose to pyruvate and arginine.35 Thus, this MG-H3 de-glycation process may have taken place during time-consuming enzymatic hydrolysis and HPLC procedures previously used for MG-H detection.37 Finally, none of the antibodies bound strongly to AGE- or CML-modified protein, consistent with findings that CML is the dominant epitope on proteins glycated with glucose in vitro.26 Lastly, we wished to investigate whether our antibodies could detect MG-Hs in a cellular system using immunofluorescence microscopy. MG-H antibodies were therefore covalently labeled with fluorophores chosen to minimize spectral overlap so that all three isomers could be concurrently visualized by triple staining. We verified by direct ELISA that this labeling did not affect epitope recognition (Data not shown). EA.hy926 cells (an immortalized human endothelial cell line38) were then treated with MGO for up to 24 h to increase MG-H formation, before being simultaneously stained with fluorescent MG-H antibodies targeting all three isomers. Immunofluorescence microscopy showed an increase in signal arising from MG-H1 and MG-H3/CEA staining with respect to the length of MGO treatment (Fig. 3). On the other hand, cells left untreated with MGO generally showed low levels of immunofluorescence (top row). Notably, fluorescence from MG-H2 antibody staining did not deviate from background levels regardless of exposure to MGO. Staining of MG-H1 and MG-H3 was observed to overlap (Fig. 4). Furthermore, although we observed some cytosolic fluorescence, most of the signal appeared to be present in the nucleus, as confirmed by co-localization with DAPI (Fig. 4). Using cells treated with MGO for varying lengths of time showed that our MG-H1 and MG-H3/CEA antibodies stained MGO-treated cells more strongly than untreated cells; however, the MG-H2 staining appeared unchanged across all conditions. This lack of change in MG-H2 staining is again consistent with

Please cite this article in press as: Wang, T.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.06.013

4

T. Wang et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx

A

CEA MG-H3 MG-H2 MG-H1 R

0.5

anti-[MG-H2] CEA MG-H3 MG-H2 MG-H1 R

1.0

Abs (450 nm)

1.0

0.0

0.5

0.0 -3

-2

-1

0

1

-3

-2

-1

[pAb] µg/mL

C

1

CEA MG-H3 MG-H2 MG-H1 R

0.6 0.4 0.2 0.0

0.0 -3

-2

-1

0

-3

1

-2

E

F

1H7G5 CEA MG-H3 MG-H2 MG-H1 R

AGEd proteins MGO-BSA MGO-HSA CML-BSA AGE-BSA BSA HSA

1.0

Abs (450 nm)

1.0

-1

[pAb] µg/mL

[pAb] µg/mL

0.5

0.0

0.5

2]

-H

-H an

an

ti-

[M

G

G

G

[M

[pAb] µg/mL

ti-

1

[M

0

ti-

-1

1]

-2

3]

0.0 -3

-H

Abs (450 nm)

0

3D11 0.8

Abs (450 nm)

Abs (450 nm)

D

CEA MG-H3 MG-H2 MG-H1 R

0.5

1

[pAb] µg/mL

anti-[MG-H3] 1.0

0

an

Abs (450 nm)

B

anti-[MG-H1]

Figure 2. Evaluation of MG-H antibodies by ELISA. (A–E) Peptides with sequence K(biotin)-LLV-X-YTKKV, X = R or MG-H/CEA modification, were bound on avidin-coated microtitre plates as the antigen. Rabbit polyclonal antibodies developed in this study against (A) MG-H1, (B) MG-H2, or (C) MG-H3 were then incubated at various dilutions, and detected using an enzyme-linked secondary antibody. Existing monoclonals (D) 3D11 and (E) 1H7G5 were evaluated for MG-H specificity in a similar fashion. (F) Recognition of various AGEd proteins by MG-H antibodies (dilution = 800 ng/mL). Error bars represent standard deviation (N = 2).

previous reports that this isomer does not form from protein glycation by MGO.11 Interestingly, although we observed some cytosolic fluorescence, the signal appeared to be strongly confined to the nuclear region. Previously reported immunofluorescence microscopy of MGO-treated endothelial cells using 1H7G5 appears to give similar patterns of staining.39 Conclusions Immunochemical reagents have proven to be invaluable in the study of protein modifications. In the case of protein glycation, anti-AGE antibodies have been used extensively to further our understanding of the fundamental biology of AGEs. Here, we have utilized our ability to chemically synthesize AGEs to generate polyclonal antibodies that recognize the three MG-H isomers. The high degree of structural similarity between these isomers has

necessitated the use of homogenous antigen for antibody purification and characterization. Furthermore, some mixed specificity might be anticipated due to the heterogeneous nature of polyclonal populations. Despite these challenges, we were pleased to discover that antibodies generated using synthetic MG-Hs were able to distinguish between the different isomers, thus highlighting the advantages of employing well-characterized antigens. Assignment of antibody specificity was conducted via a direct ELISA employing synthetic MG-H peptides. The development of this assay also allowed us to characterize two previously reported monoclonal antibodies thought to recognize MG-H modifications.32,34 Interestingly, both antibodies showed cross-reactivity between two or more MGO-derived species. Clone 3D11 recognizes MG-H1 and MG-H3 with near-equal affinity, and also strongly cross-reacts with CEA. On the other hand, 1H7G5 shows equivalent binding to MG-H3 and CEA. This latter observation, along with the

Please cite this article in press as: Wang, T.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.06.013

T. Wang et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx

5

Figure 3. Immunofluorescence microscopy of cells treated with MGO and triply stained with MG-H antibodies. EA.hy926 cells treated with 250 lM MGO (+MGO) for 1, 5, or 24 h show increased staining with fluorescently labeled MG-H1 and MG-H3/CEA antibodies compared to cells not treated with MGO (neg), with fluorescence intensity increasing with MGO incubation time. In contrast, MG-H2 staining showed no differences in staining regardless of treatment condition. Exposure settings were adjusted to avoid saturated pixels in the brightest objects and were held constant across all experiments. Scale bars represent 50 lm.

Figure 4. Immunofluorescence microscopy of MGO treated cells double-labeled with MG-H antibodies. EA.hy926 cells treated with 250 lM MGO (+MGO) for 24 h show increased staining with fluorescently labeled MG-H1 and MG-H3/CEA antibodies over untreated cells (neg). Localization of staining appears to be predominantly nuclear, as evidenced by DAPI co-localization. Scale bars represent 50 lm.

binding preferences of our own MG-H3 antibody, is surprising given the structural dissimilarities between MG-H3 and CEA, and

raises several intriguing possibilities. For example, we cannot rule out that structural features in the antibody-combining site that

Please cite this article in press as: Wang, T.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.06.013

6

T. Wang et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx

enable binding to MG-H3 might also impart CEA recognition as well. Indeed, monoclonal antibodies have been previously reported to cross-react with multiple unrelated antigens,40 a property termed ‘bi-specificity.’ Alternatively, these antibodies may be ‘catalytic’ in nature, and accelerate the chemical interconversion between MG-H3 and CEA. Studies to differentiate between these and other possibilities are currently ongoing in our laboratory. Immunofluorescence experiments with MG-H antibodies revealed higher levels of MG-H1 and MG-H3/CEA formation in MGO-treated cells. The staining observed was predominantly nuclear in localization. Although MGO should freely diffuse into the nucleus due to its small size,41 due to its reactivity, we expected a more homogenous distribution of MG-H modifications within the cell. However, argpyrimidine, another MGO-arginine AGE, has also been observed to localize in the nucleus.39 Thus, it is possible that some nuclear proteins may be especially susceptible to MGO modification, or that proteins may concentrate in the nucleus upon MGO modification. However, we cannot rule out a scenario where these antibodies may be cross-reacting to some extent with MGO-modified nucleobases, some of which have structural similarity to the MG-Hs or CEA.12,42 In summary, our ability to access to chemically homogeneous MG-Hs has allowed for the development of polyclonal antibodies with specificity most likely unachievable through the use of traditional immunogens. These reagents should prove to be useful for the visualization of MG-H modifications in biological samples. Furthermore, the strategy employed herein should enable the development of well-characterized antibodies targeting a wide range of AGE epitopes. Acknowledgements We would like to express our gratitude to the SENS foundation for generous financial support, and Drs. Aubrey de Grey and William Bains for helpful discussions. We wish to thank Professor Michael Brownlee for the generous gift of antibody 1H7G5 and Scott Lewis of NEP for helpful discussions. D.A.S. is a paid consultant for Bristol-Myers Squibb. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.bmcl.2015.06. 013. References and notes 1. Singh, R.; Barden, A.; Mori, T.; Beilin, L. Diabetologia 2001, 44, 129. 2. Hellwig, M.; Henle, T. Angew. Chem. Int. Ed. 2014, 53, 10316. 3. Sell, D. R.; Biemel, K. M.; Reihl, O.; Lederer, M. O.; Strauch, C. M.; Monnier, V. M. J. Biol. Chem. 2005, 280, 12310. 4. Makita, Z.; Radoff, S.; Rayfield, E. J.; Yang, Z.; Skolnik, E.; Delaney, V.; Friedman, E. A.; Cerami, A.; Vlassara, H. N. Engl. J. Med. 1991, 325, 836.

5. Heijst, J. W.; Niessen, H. W.; Hoekman, K.; Schalkwijk, C. G. Ann. N. Y. Acad. Sci. 2005, 1043, 725. 6. van Heijst, J. W.; Niessen, H. W.; Musters, R. J.; van Hinsbergh, V. W.; Hoekman, K.; Schalkwijk, C. G. Cancer Lett. 2006, 241, 309. 7. Kilhovd, B. K.; Berg, T. J.; Birkeland, K. I.; Thorsby, P.; Hanssen, K. F. Diabetes Care 1999, 22, 1543. 8. Simm, A.; Wagner, J.; Gursinsky, T.; Nass, N.; Friedrich, I.; Schinzel, R.; Czeslik, E.; Silber, R.; Scheubel, R. Exp. Gerontol. 2007, 42, 668. 9. Yamagishi, S.; Nakamura, K.; Inoue, H.; Kikuchi, S.; Takeuchi, M. Med. Hypotheses 2005, 64, 1205. 10. Meerwaldt, R.; Links, T.; Zeebregts, C.; Tio, R.; Hillebrands, J.-L.; Smit, A. Cardiovasc. Diabetol. 2008, 7, 29. 11. Klopfer, A.; Spanneberg, R.; Glomb, M. A. J. Agric. Food Chem. 2011, 59, 394. 12. Thornalley, P. J. Drug Metab. Drug Interact. 2008, 23, 125. 13. Kilhovd, B. K.; Juutilainen, A.; Lehto, S.; Rönnemaa, T.; Torjesen, P. A.; Hanssen, K. F.; Laakso, M. Atherosclerosis 2009, 205, 590. 14. Ahmed, N.; Ahmed, U.; Thornalley, P. J.; Hager, K.; Fleischer, G.; Münch, G. J. Neurochem. 2005, 92, 255. 15. Kilhovd, B.; Giardino, I.; Torjesen, P.; Birkeland, K.; Berg, T.; Thornalley, P.; Brownlee, M.; Hanssen, K. Metab.: Clin. Exp. 2003, 52, 163. 16. Han, Y.; Randell, E.; Vasdev, S.; Gill, V.; Curran, M.; Newhook, L. A.; Grant, M.; Hagerty, D.; Schneider, C. Clin. Biochem. 2009, 42, 562. 17. Fosmark, D. S.; Torjesen, P. A.; Kilhovd, B. K.; Berg, T. J.; Sandvik, L.; Hanssen, K. F.; Agardh, C.-D.; Agardh, E. Metab.: Clin. Exp. 2006, 55, 232. 18. Westwood, M. E.; Thornalley, P. J. J. Protein Chem. 1995, 14, 359. 19. Yanagisawa, K.; Makita, Z.; Shiroshita, K.; Ueda, T.; Fusegawa, T.; Kuwajima, S.; Takeuchi, M.; Koike, T. Metab.: Clin. Exp. 1998, 47, 1348. 20. Ahmed, N.; Thornalley, P. J. Biochem. J. 2002, 364, 15. 21. Ahmed, N.; Argirov, O. K.; Minhas, H. S.; Cordeiro, C. A.; Thornalley, P. J. Biochem. J. 2002, 364, 1. 22. Bansode, S. B.; Chougale, A. D.; Joshi, R. S.; Giri, A. P.; Bodhankar, S. L.; Harsulkar, A. M.; Kulkarni, M. J. Mol. Cell. Proteomics: MCP 2013, 12, 228. 23. Gomes, R. A. FEMS Yeast Res. 2008, 8, 174. 24. Ikeda, K.; Higashi, T.; Sano, H.; Jinnouchi, Y.; Yoshida, M.; Araki, T.; Ueda, S.; Horiuchi, S. Biochemistry 1996, 35, 8075. 25. Schleicher, E. D.; Wagner, E.; Nerlich, A. G. J. Clin. Invest. 1997, 99, 457. 26. Reddy, S.; Bichler, J.; Wells-Knecht, K. J.; Thorpe, S. R.; Baynes, J. W. Biochemistry 1995, 34, 10872. 27. Ahmed, M.; Frye, E. B.; Degenhardt, T.; Thorpe, S.; Baynes, J. Biochem. J. 1997, 324, 565. 28. Nagai, R.; Fujiwara, Y.; Mera, K.; Yamagata, K.; Sakashita, N.; Takeya, M. J. Immunol. Methods 2008, 332, 112. 29. Taneda, S.; Monnier, V. M. Clin. Chem. 1994, 40, 1766. 30. Oya, T.; Hattori, N.; Mizuno, Y.; Miyata, S.; Maeda, S.; Osawa, T.; Uchida, K. J. Biol. Chem. 1999, 274, 18492. 31. Takeuchi, M.; Yanase, Y.; Matsuura, N. Mol. Med. 2001, 7, 783. 32. Cai, W.; Gao, Q. D.; Zhu, L.; Peppa, M.; He, C.; Vlassara, H. Mol. Med. 2002, 8, 337. 33. Nishikawa, T.; Edelstein, D.; Du, X. L.; Yamagishi, S.; Matsumura, T.; Kaneda, Y.; Yorek, M. A.; Beebe, D.; Oates, P. J.; Hammes, H. P.; Giardino, I.; Brownlee, M. Nature 2000, 404, 787. 34. Giardino, I.; Thornalley, P.; Edelstein, D.; Brownlee, M. In Diabetes; AMER DIABETES ASSOC: 1660 DUKE ST, ALEXANDRIA, VA 22314 USA, 1998; Vol. 47, p A123. 35. Wang, T.; Kartika, R.; Spiegel, D. A. J. Am. Chem. Soc. 2012, 134, 8958. 36. Nagai, R.; Mera, K.; Nakajou, K.; Fujiwara, Y.; Iwao, Y.; Imai, H.; Murata, T.; Otagiri, M. Biochim. Biophys. Acta 2007, 1772, 1192. 37. Ahmed, N.; Dobler, D.; Dean, M.; Thornalley, P. J. J. Biol. Chem. 2005, 280, 5724. 38. Edgell, C. J.; McDonald, C. C.; Graham, J. B. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 3734. 39. Nakadate, Y.; Uchida, K.; Shikata, K.; Yoshimura, S.; Azuma, M.; Hirata, T.; Konishi, H.; Kiyama, H.; Tachibana, T. Biochem. Biophys. Res. Commun. 2009, 378, 209. 40. James, L. C.; Roversi, P.; Tawfik, D. S. Science 2003, 299, 1362. 41. Bagley, S.; Goldberg, M.; Cronshaw, J.; Rutherford, S.; Allen, T. J. Cell Sci. 2000, 113, 3885. 42. Li, Y.; Cohenford, M. A.; Dutta, U.; Dain, J. A. Anal. Bioanal. Chem. 2008, 390, 679.

Please cite this article in press as: Wang, T.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.06.013