Background of the Invention

The present invention relates to a process for isolating and identifying a novel imidazo [4,5b] pyridinium molecule, referred to by the inventors as "pentosidine", from the extracellular matrix of humans and other mammals. In addition, the invention is directed to the production of antibodies against pentosidine and for the use of the antibodies to detect the molecule in tissues, fluid samples, etc. Furthermore, the invention relates to the quantitation of pentosidine in tissues in order to assess aging in humans and other mammals.

The recently isolated imidazo [4,5b] pyridinium molecule, or pentosidine, is believed to be produced according to the non-enzymatic reaction of sugars with various amino acid or protein residues during the aging and/or degradation of proteins. In this regard, the pentosidine molecule has been structurally characterized to consist essentially of a lysine and an arginine residue crosslinked by a pentose. Furthermore, the novel imidazo [4,5b] pyridinium or pentosidine molecule of the invention has been chemically synthesized by the inventors in order to confirm the structural arrangement of the isolated

SUBSTITUTESHEET(RULE25)

molecule. The present invention is further directed to the use of the recently isolated, characterized, and chemically synthesized pentosidine molecule and/or antibodies thereto in various processes and/or compositions for studying the aging and/or degradation of proteins in humans and other mammals.

The extracellular matrix of humans and other mammals undergoes progressive changes during aging that are characterized by decreased solubility (Schnider, S.L., and Kohn, R.R. , J. Clin. Invest. 67, pp. 1630-1635, 1981), decreased proteolytic digestibility (Hamlin, C.R. , Luschin, J.H. , and Kohn, R.R. , EXP. Gerontol. 13, pp. 415-523, 1978), increased heat denaturation time (Snowden, J.M. , Eyre, D.R., and Swann, D.H. , Biochem. Biophvs. Acta. 706, pp. 153-157, 1982) and accumulation of yellow and fluorescent material (LaBella, F.S., and Paul, G. , J. Gerontol.. 20, pp. 54-59, 1964) . These changes, which affect particularly collagen-rich tissues and appear to be accelerated in diabetes, are thought to result from the formation of age-related intermolecular crosslinks.

Elucidation of the structure of these age-related intermolecular crosslinks has been for many years of major interest to gerontologists and collagen chemists for two principal reasons. First, there appears to exist an inverse relationship between mammalian longevity and aging rate of collagen (Kohn, R.R. in Testing the Theories of

Aging (Adelman, R.C., and Roth, G.S., eds.) pp. 221-231, CRC Press, Inc., Boca Raton, Florida) suggesting that the process which governs longevity may express itself at least partially in the aging rate of collagen. Second, the progressive increase in stiffness of collagen-rich tissues like arteries, lungs, joints and the extracellular matrix has been associated with age-related diseases such as hypertension, emphysema, decreased joint mobility and ability to fight infections. Thus, elucidation of the nature of extracellular matrix crosslinking in aging is of both practical and theoretical interest.

Along these lines, the present inventors and others previously postulated that the advanced Maillard or nonenzymatic glycosylation reaction which occurs between reducing sugars, e.g., glucose, and amino groups on proteins could explain some of the age and diabetes-related changes that affect long-lived proteins through browning and crosslinking (Monnier, V.M. , and Cerami, A., Science. 211, pp. 491-493, 1981). However, direct demonstration of this hypothesis has not been possible since the structures of Maillard protein adducts and crosslinks were previously unknown.

In this regard, Cerami, et al. , U.S. Patent Nos. 4,665,192 and 4,758,583 reported the discovery of a new and useful fluorescent chro ophore -2-(2-furoyl)-4(5)- 2(furanyl)-lH-imidazole (FFI) and a method of utilizing

SUBSTITUTESHEET(RULE25)

this chromophore for inhibiting protein aging. However, the present inventors have demonstrated that the FFI compound described in these patents is merely an artifact of acid hydrolysis and alkalization with ammonia and is not one of the end products of extended non-enzymatic polypeptide glycosylation (Njoroge, et al., J.Biol.Chem.. 263: 10646-10652, 1988).

However, notwithstanding the above, recent observations continue to suggest that some of the changes occurring in the aging process of collagen could be explained by the Maillard or non-enzymatic browning reaction which occurs in stored or heated foodstuffs (Monnier, V.M. and Cerami, A., Am.Chem.Soc. 215, 431, 1983) . In this regard, reducing sugars react non- enzymatically with the free amino groups of the proteins to form insoluble, highly crosslinked, yellow and fluorescent products. Studies on the potential occurrence of the non- enzymatic browning reaction in vivo demonstrated an age- related increase in dura and skin collagen-linked fluorescence at 440 nm (excitation at 370 nm) and chromophores absorbing above 300 nm (Monnier, V.M. , Kohn, R.R. , and Cerami, A., Proc. Natl. Acad. Sci. 81, 583, 1984) (Monnier, V.M. , Vishwanath, v., Frank, K.E., Elmets, CA. , Dauchot, P., and Kohn, R.R. , New Enσl. J. Med. 314, 403, 1986) . Similar spectroscopical changes could be duplicated by incubating collagen with reducing sugars such as

SUBSTITUTESHEET(RULE25)

glucose, glucose-6-phosphate or ribose (Monnier, V.M. , Kohn, R.R., and Cerami, A., Proc. Natl. Acad. Sci. 81, 583, 1984) (Kohn, R.R. , Cerami, A., Monnier, V.M. , Diabetes 33, 57, 1984) . In addition, it was demonstrated that collagen incubated with these sugars was highly crosslinked suggestingthat the sugar-derived fluorophores-chromophores could act as intra- or intermolecular crosslinks (Monnier, V.M. , Kohn, R.R. , and Cerami, A., Proc. Natl. Acad. Sci. 81, 583, 1984) (Kohn, R.R. , Cerami, A., Monnier, V.M. , Diabetes 33, 57, 1984).

The potential role of the Maillard reaction in these changes was further substantiated by the observation that non-enzymatic glycosylation which initiates the Maillard reaction was increased in diabetic and aging collagen and by observations in subject with Type I (insulin-dependent) diabetes that revealed a dramatic increase in collagen- linked fluorescence (Monnier, V.M. , Vishwanath, V., Frank, K.E., Elmets, C.A., Dauchot, P., and Kohn, R.R. , New Enσl. J. Med. 314, 403, 1986) (Vishwanath, V., Frank, K.E., Elmets, CA. , Dauchot, P.J., Monnier, V.M. , Diabetes 35, 916, 1986).

Although age-related acceleration of collagen browning may be explained by the Maillard reaction, the evidence presented for support of this hypothesis has been very circumstantial. More particularly, such evidence is based on spectroscopical changes of collagen with aging and

SUBSTITUTESHEET(RULE25)

diabetes in vivo with conspicuous similarities produced by the incubation of young collagen with reducing sugars in vitro. Because of uncertainty in the exact nature of the fluorescence produced during the aging of proteins, as well as the particular nature of the protein adducts and crosslinks involved therein, the present inventors initiated a study that resulted in the present invention with the ultimate aim of elucidating the nature of the collagen-linked fluorescence which increases in aging and diabetes.

In this regard, the present inventors conducted a systematic investigation of the chemical nature of the fluorescence that accumulates in aging human collagen. Two novel fluorophores, nicknamed "P" and "M", with excitation/emission maxima at 335/385 nm and 360/460 n , respectively, were isolated from insoluble collagen following proteolytic digestion and chromatography (Sell, D.R. , and Monnier, V.M. , Conn. Tiss. Res. 19, pp. 77-92, 1989) . An age-related effect was noted for both types of fluorophores (i.e. the presence of the fluorophores increased with age) . Although fluorophore M was borohydride reducible and unstable to acid hydrolysis, thereby suggesting that M had an iminopropene-type configuration which substantiated, but did not prove, that glucose was responsible for its origin, the fluorescence properties of the 335/385 fluorophore, i.e. fluorophore

SUBSTITUTESHEET(RULE25)

"P", were found unchanged following acid hydrolysis in 6 N HC1 for 24 hours at 110°C As a result of its resistance to acid hydrolysis, a larger quantity of fluorophore P was purified from acid hydrolyzed dura mater collagen and its structure was elucidated using H-NMR, ^I3 C-NMR and MS/MS fast atom bombardment spectroscopy. Structure elucidation of fluorophore "P" led to the discovery of a pentose- mediated protein crosslink named "pentosidine".

Further, the present inventors have developed antibodies specifically immunoreactive to pentosidine, and particularly immunoreactive to synthetic pentosidine. These antibodies are useful in determining the presence of pentosidine in biological samples.

In addition, the inventors have found that levels of pentosidine progressively increase with age in various tissues of human origin, including dura mater, skin and cartilage. Levels in skin correlate with severity of complication in individuals with insulin-dependent diabetes and are elevated during end-stage renal disease. Although originally isolated from the highly insoluble fraction of collagen-enriches dura mater, pentosidine has also been identified in various other tissues and proteins.

One of the tests for the usefulness of a parameter as an indicator of aging is that the aging rate must be accelerated in species that have short life spans in comparison with those of longer life spans. The

SUBSTITUTESHEET(RULE25)

inventors have examined the levels of pentosidine in skin collagen in various animal species of differ life spans in order to determine the potential usefulness of pentosidine as a marker of a fundamental aging process.

Summary of the Invention

In one aspect, the present invention is directed to a novel imidazo [4,5b] pyridinium molecule composed of a lysine and an arginine residue crosslinked with a pentose sugar. The novel imidazo [4,5b] pyridinium compound, referred to as "pentosidine", was isolated from proteineous tissue undergoing advanced glycosylation and is believed to be one of the principal products involved in the non- enzymatic browning and/or aging of proteins. Assaying for the pentosidine molecule makes it possible to assess the degree of non-enzymatic glycosylation occurring. In addition, the pentosidine molecule may be utilized through the production of antibodies thereto and/or the preparation of test kits, etc. for diagnostic, as well as therapeutic purposes (i.e. development of agents which inhibit the non- enzymatic browning reaction, etc.).

Structural elucidation of the pentosidine molecule indicates that its precise chemical name 3-H-imidazole

[4,5b] pyridine-4-hexanoic acid, alpha amino-2 [ (4-amino-4- carboxybutyl) amino] , and that its structural composition is as follows:

SUBSTITUTESHEET(RULE25)

PENTOSONE

In a further aspect, the present invention is directed to a process for chemically synthesizing the pentosidine molecule. The structure of the isolated pentosidine molecule was confirmed by the non-enzymatic reaction of ribose with lysine and arginine residues.

In additional aspect, the present invention is directed to a process for isolating the pentosidine molecule from insoluble collagen tissue through the acid- hydrolysis of insoluble collagen and the structural elucidation of the isolated molecule using ^! H-NMR, ^I3 C-NMR and various other spectroscopy techniques.

In a still further aspect, the present invention provides for the generation of immunoreactive antibodies against the pentosidine molecule or parts thereof, including synthetic molecules. These antibodies are useful for specifically detecting the pentosidine molecule in vitro and in vivo.

In a still additional aspect, the invention relates to a method of quantitating pentosidine in tissue in order to assess aging in mammals. The inventors have found that an

SUBSTITUTE SHEET(RULE25)

inverse relationship is present between a mammal's longevity and the glycoxidation rate of pentosidine formation. Similarly, the present invention is also directed to a process of enhancing longevity of a mammal by decreasing the rate of pentosidine produced through glycoxidation.

Other aspects and advantages of the present invention will become apparent to those skilled in the art upon a review of the following materials.

Brief Description of the Drawings

The following is a brief description of the drawings which are presented for the purposes of illustrating the invention and not for the purposes of limiting same.

FIGURES 1A-1C are graphs illustrating the results [i.e. absorbance (Figure 1A) , hydroxyproline (Figure IB), and fluorescence (Figure 1C) ] , produced by the fractions obtained by gel filtration on Bio-Gel P-2 of acid- hydrolyzed human dura mater. Fluorescence was monitored at excitation/emission wavelengths of 335/385 nm. FIGURES 2A and 2B are graphs illustrating the absorption (Figure 2A) and fluorescence-excitation (Figure 2B) spectra at pH 2, 7, 9, and 12 of fluorophore P (pentosidine) isolated from native human dura mater (upper) and a synthetic incubation system of heating lysine, arginine, and ribose together at 80°C for one hour (lower) .

SUBSTITUTESHEET(RULE25)

Fluorescence-excitation spectra were monitored as follows: for the emission spectra on the right, excitation was at 335 nm, and for the excitation spectra on the left, emission was at 385 nm. FIGURE 3 is a Η-NMR spectra of pentosidine isolated from human dura mater and a synthetic incubation system.

FIGURE 4 is a expanded 7.1 - 8.2 portion of the Η-NMR spectrum from FIGURE 3 (native) .

FIGURE 5 is a graph showing a comparison of a HPLC chromatogram of purified pentosidine (P) from native dura mater to that of unpurified synthetic material. Separations were made on a 0.46 x 25 cm Vydac 218 TP C-18 column by application of a linear gradient of 10-17% acetonitrile in water applied from 0-35 minutes at a flow rate of 1 ml/min. with 0.01 M HFBA as the counterion. Fluorescence was monitored as of Figure 1.

FIGURE 6 is a FAB CAD MS/MS spectra of pentosidine isolated from human dura mater and a synthetic incubation system. FIGURE 7 is a chart indicating the proposed mechanism for the formation of pentosidine.

FIGURE 8 is a graph showing the relationship of the pentosidine level as a function of age in human dura mater.

The assay was conducted according to conditions of Figure 5. Line equation: pentosidine (pmol/mg collagen)= 10.8 +

2.43 (age). N=24, p<.001, r=91.

SUBSTITUTE 5HEET (RULE 25)

FIGURE 9 is a graph showing the relationship of the pentosidine level as a function of age in human skin. Regression line equation: y = 0.002X ² + 0.214X + 5.69.

FIGURE 10 is a graph showing the relationship between pentosidine and the presence of diabetes or renal failure.

Levels are expressed relative to a normal range computed as a 95% confidence interval determined and reproduced from the regression line of Figure 9.

FIGURE 11 are two graphs showing the distribution of pentosidine values in acid hydrolysate of plasma and hemolysate from control (CON) , diabetic (DB) , and uremic (end stage renal disease [ESRD]) plasma obtained by combined reverse-phase high-performance liquid and ion- exchange chromatography. FIGURE 12 is a bar graph showing age-adjusted pentosidine ± SE in diabetic skin versus the cumulative sum of all diabetic complications, as established by combining all indexes of severity ranging from 0 to 8 for complications listed in Table IV. FIGURE 13 is a drawing showing various compounds tested for cross-reactivity with antibody to pentosidine. These compounds are identified as 1, Pentosidine; 2, azabenzimidazole; 3, benzimidazole; 4, quinoline; 5, ademine; 6, guanine; 7, /3-carboline; 8, pyridine; 9, pyrimidine; 10, quanidine; 11, L-arginine; 12, L-lysine, R _j , arginyl; R ₂ , lysyl residue.

SUBSTITUTESHEET(RULE25)

FIGURE 14 is a graph illustrating the time course of anti-pentosidine antibody titer in rabbit serum immunized with KLH-pentosidine conjugate. Titer in blood from the rabbit with the highest titer, drawn before immunization (Δ) and at 2 (■) , 4 (T) , 7 (O) , 9 (v) , and 13 weeks (•) after immunization.

FIGURE 15 is a graph showing the enzyme-linked im unoassay of free pentosidine. Curve represents inhibition of antibody binding by free pentosidine. B, A _w in the presence of competitior; B ₀ , A ₀₅ in the absence of competitiors. Bars show + SD (n=6) .

FIGURE 16 is a graph illustrating the time course of pentosidine formation in BSA incubated with 100 mmol/L ribose at 37°C for various times: 0 (O) , 3 (•) , 5 (v) , 7 (T) , 14 (O) , 30 (■), 60 (Δ) , and 90 days (A). Lowest curve

(Δ) is dilution curve for pentosidine.

FIGURE 17A-17C are SDS-Page prints illustrating (A) SDS-PAGE of lysozyme incubated with ribose for various periods and stained with Coornassie blue; (B) immunoblot of lysozyme-ribose with anti-pentosidine antibody; (C) immunoblot of lysozyme-ribose with anti-pentosidine antibody preincubated with free pentosidine. Lane 1, unmodified lysozyme; lane 2, incubated for 1 day; lane 3, 4 days; lane 4, 8 days; lane 5, 11 days; lane 6, 15 days; and lane 7, prestained standard (low-molecular-mass, Bio- Rad) . M.D. and T. monomer, dimer, and trimer, respectively.

SUBSTITUTESHEET(RULE25)

FIGURE 18 are two graphs illustrating digestibility and immunoreactivity of plasma proteins incubated with 20 g/kg Pronase E at 37°C for various times. (Lower panel)

Ninhydrin-reactive amino groups in enzyme-digested pool plasmas with low (π) , middle (v) , and high (O) pentosidine concentrations compared with those in acid-hydrolyzed plasmas. (Upper panel) Change of immunoreactivity of enzyme-digested pool plasmas with low (Q) , middle (Δ) , and high (O) pentosidine concentrations compared with that of acid-hydrolyzed plasmas (■, r, •) .

FIGURE 19 is a graph showing a comparison of immunoreactivity curves of various compounds with anti- pentosidine antibody calibration curve; Pronase E digested BSA-ribose (O) , Pronase E-digested lysozyme-ribose (■) , pronase E-digested diabetic plasma (□) , pentosidine (•) . FIGURE 20 is a graph illustrating the correlation between pentosidine in normal (Δ: inset), diabetic (O) , and uremic (O) plasma measured by ELISA and HPLC

FIGURE 21 is a graph showing the plasma pentosidine (mean + SD) in normal subjects, patients without diabetes or uremia (non-DB, non-U) , diabetic patients (DB) , and uremic patients. Values in diabetic and uremic subjects are significantly greater than in controls (P<0.01 and P<0.0001, respectively). FIGURE 22 is a graph illustrating the correlation between age and pentosidine concentrations in skin from

SUBSTITUTESHEET(RULE25)

normal (o, diabetic (•), and uremic (r) subjects and diabetic subjects with uremia (■) . The regression line and 95% confidence interval are shown.

FIGURES 23A-23F are graphs showing the relationship of pentosidine level in skin collagen as a function of age in different animal species. Shown are the regression line and 95% confidence interval of predicted values.

FIGURE 24 is a composite graph summarizing 95% confidence intervals of predicted values (as indicated by lines and shaded areas) for animal species of FIGURE 23.

FIGURES 25A-25C are graphs illustrating the relationship of pentosidine as a function of age in skin biopsies from ad libitum-fed squirrel or rhesus monkey.

(A) squirrel monkey, y=0.89x - 3.0, R=0.84, n=15, age (P<.0004); (B) rhesus monkey, y=0.23x + 3.7, R=0.75, n=24, age (P<.015); (C) 95% confidence intervals of predicted values per monkey species as indicated, (π) : level from monkey with a single biopsy taken in either one of the two time periods; (O) : level from monkey with two biopsies taken two years apart.

FIGURES 26A-26C are graphs showing the effects of age and dietary restriction on (A) tendon breaking time, (B) tendon pentosidine and (C) skin pentosidine levels in Fischer 344 rats. Error bars + standard deviation. Means that do not share the same superscript are significantly different (P<0.05).

SUBSTITUTESHEET(RULE25)

FIGURES 27A-27B are graphs demonstrating the rate of pentosidine formation in skin shown as plot of coefficients of x ² from Table 2 versus: (A) maximum life span and (B) logarithm of maximum life span. Line equations: (A) 10*=2.917X- ⁰ - ²⁵² , (B) y=0.465 - 0.252X' where x'=logx, R=- 0.854, n=9, p<0.0035.

Detailed Description of the Invention The present invention is directed to the isolation and identification of an acid resistant fluorescent molecule from the extracellular matrix of humans and other mammals. Structure elucidation of the isolated fluorescent molecule revealed the presence of an imidazo [4,5b] pyridinium molecule comprising a lysine and an arginine residue crosslinked by a pentose. Confirmation of this structural arrangement was achieved in vitro by the non-enzymatic reaction of ribose with lysine and arginine residues.

In addition, it has been determined that the newly discovered crosslink, named "pentosidine" by the inventors, can also be synthesized with isomers of ribose, arabinose, xylose, and lyxose, as well as by incubating young human collagen with these sugars at 37°C Moreover, pentosidine was found in a variety of human tissues including plasma proteins and red blood cells. Its presence in cells grown in culture strongly suggests ribose or ribonucleotide

SUBSTITUTESHEET(RULE25)

metabolites as precursors. The unexpected discovery of pentose-mediated protein crosslinking, as well as the pentosidine crosslink, provides useful tools for the further investigation and explanation of the aging process. More particularly, the present invention relates to the use of a novel fluorophore compound (formerly referred to as "flurophore P" and now referred to as "pentosidine") which has been isolated from human collagen undergoing advanced non-enzymatic glycosylation and identified as 3-H- imidizol [4,5b] pyridine-4-hexanoic acid, alpha amino-2[ (4- amino-4-carboxybutyl) amino] . The imidazopyridinium compound is believed to be one of the end products of the extended non-enzymatic polypeptide glycosylation reaction normally associated with the structural and functional changes in tissues that occur during the aging process, and has also been observed to occur at an accelerated rate in individuals suffering from diabetes. By identifying the occurrence of advanced glycosylation through the detection of the specified fluorophore compound of the present invention, the degree of cellular stress or injury caused by diabetes, aging, and/or uremia may be determined. In addition, detection of the pentosidine compound may also aid in determining who among diabetic subjects is at risk of developing diabetic complications. Thus, the newly discovered fluorescent imidazopyridinium compound, or pentosidine, as well as antibodies specific to said

SUBSTITUTESHEET(RULE25)

compound, may be used in connection with various diagnostic techniques, to determine the advancement of glycosylation in protein specimens.

Additionally, since it is generally thought that the aging effects produced by the non-enzymatic polypeptide glycosylation of the significant protein masses of the body (such as collagen, elastin, lens protein, nerve proteins, and the kidney glo erular basement membranes) is caused by the cross-linking of sugars with the amino acids of the proteins, the imidazopyridinium compound of the present invention (i.e. pentosidine) may also be utilized as an exploratory tool for the development and testing of possible agents capable of interfering with the cross¬ linking process, thereby inhibiting protein aging. Hence, the present invention may reduce the incidence of pathologies involving the cross linking of proteins such as atherosclerosis, osteoarthritis, loss of elasticity and wrinkling of the skin, and stiffening of joints.

The present inventors have developed a process for isolating and purifying a 2-alkyl amino-4-alkyl imidazopyridinium compound (specifically, 3-H-imidazole [4,5b] pyridine-4-hexanoic acid, alpha amino-2[ (4-amino-4- carboxybutyl) amino], i.e. "pentosidine"), a newly discovered imidazopyridinium compound representing a cross link between the amino acids lysine and arginine, from a pool of insoluble human dura mater collagen following

SUBSTITUTESHEET(RULE25)

enzymatic hydrolysis and sequential purification steps utilizing Sephadex G50, paper, cation, and high performance liquid chromatography (the specific procedures and material involved in this process are more clearly set forth below in the examples) . The imidazopyridinium compound was detected on the basis of its fluorescence at 385 nm upon excitation at 335 nm. Its maximum UV was at 325 nm.

Furthermore, since the compound was not destroyed by acid hydrolysis, this allowed it to be directly prepared from batch quantities of collagen. More particularly, because the pentosidine molecule was resistant to acid hydrolysis, a larger quantity of the molecule could be purified from acid hydrolyzed dura mater collagen. The compound may then be assayed by HPLC with a fluorescence detector. Structure elucidation of the fluorescent compound by Η-NMR, COSY, ¹³ C-NMR, MS/MS FAB Spectroscopy indicated the presence of an imidazopyridinium compound involving lysine and arginine and a 5-carbon moiety in the heterocyclic ring. Consequently, pentosidine, a cross-link consisting of arginine, lysine, and a pentose sugar and one of the advanced products of the Maillard reaction, has been isolated and characterized. Since pentosidine's mechanistic formation involves both glycation and oxidation, the term "glycoxidation product" has been

SUBSTITUTESHEET(RULE25)

applied to delineate this novel Maillard product resulting from oxidative cleavage of A adori products.

Moreover, in a study involving skin specimens obtained at autopsy, the imidazopyridinium compound (pentosidine) was found to increase exponentially with age. High levels were detected in diabetic subjects with nephropathy as well as in uremic subjects without nephropathy.

In addition, the inventors have chemically synthesized their imidazopyridinium compound by the following general procedure. A mixture consisting of 0.1 M D-ribose, L- lysine HC1 and L-arginine at pH 7.4 was heated to 80°C for 60 minutes and passed over Dowex 50 x 4 resin (H ⁺ -form) . The resin was washed with one liter H ₂ 0, 1 M pyridine with 1L of 2N NaOH. After neutralization and evaporation the concentrate was chromatographed over Bio-Gel P-2 equilibrated in 0.02 M Hepes. Fractions containing the fluorophore were further purified by HPLC using a C-18 reverse phase column and a linear gradient of acetonitrile with 0.01M of heptafluorobutyric acid (HFBA) as counterion. All fluorescence, UV, NMR, and mass spectroscopical properties of the synthetic compound were rigorously identical with those of the native compound. Yield was 0.1% under non-optimized reaction conditions.

The possibility of producing the imidazo-pyridinium compound with pentoses came to the inventors as a big surprise. In this regard, the use of pentoses (i.e.

SUBSTITUTESHEET(RULE25)

ribose, arabinose, and xylose) in the total synthesis of pentosidine or any imidazopyridinium compound derived from the general reaction:

whereby R = aliphatic or aromatic rest.

is completely new. As more particularly discussed below, the in vitro synthesis of pentosidine and/or other imidazopyridine compounds is a valuable tool for researching the aging process.

Furthermore, antibodies specifically immunoreactiveto pentosidine have also been produced. The antibodies detected by ELISA reacted strongly with free pentosidine but not with pentosidine-like compounds. The working range of the competitive ELISA for standard pentosidine was 0.1-100 pmol. Pentosidine was detectable in bovine serum albumin incubated with ribose as a function of incubation time. Immunoblotting studies showed that pentosidine specifically stained in oligomers of lysozyme incubated with ribose. Digestion with protease (Pronase E, 20 g/kg) as well as acid hydrolysis enhanced the immunoreactivity of samples, the pentosidine values in digested human plasma correlating with those measured by HPLC (r = 0.88) . Pentosidine in diabetic and uremic plasma digested with

SUBSTITUTE SHEET(RULE25)

Pronase E was significantly higher than normal (P<0.01; mean ± SD) : 1620 ± 1940 and 2630 ± 1360 nmol/L, respectively, vs 151 + 55 nmol/L (normal) . Amounts of pentosidine in hydrolyzed skin collagen increased with age and were increased in diabetes and uremia. This ELISA provides a new tool for assessing the role of the advanced Maillard reaction in aging and age-related diseases.

Additionally, the present invention is directed to a process for assessing aging on a mammal characterized by the steps of quantitating the amount of pentosidine produced by glycoxidation which is present in the mammal's tissue such as skin collagen. The inventors have found that an inverse relationship exists between the mammals longevity and the glycoxidation rate of pentosidine formation.

Similarly, the rpesent invention relates to a process of enhancing longevity of a mammel. This is through the steps of decreasing the rate of pentosidine formation through glycoxidation. The specific procedures and materials used in the isolation, quantitation, characterization, and chemical synthesis of the pentosidine compound of the present invention and/or the production of antibodies to the isolated and chemically synthesized molecules are set forth below in the following illustration examples.

SUBSTITUTESHEET(RULE25)

Example 1 METHODS AND PROCEDURES

Preparation of Pentosidine from Dura Mater

The starting material consisted of insoluble human dura mater (60 g. wet weight) determined to contain greater than 95% collagen on the basis of hydroxyproline content. Hydroxyproline was quantitated as described in Hamlin, et al. (i.e. Hamlin, C.R. , Luschin, J.H. , and Kohn, R.R., EXP. Gerontol. 13, pp. 415-523, 1978) and assumed to make up 14% of the collagen by weight (Hamlin, C.R. , and Kohn, R.R., Biochem. Biophvs. Acta. 236, pp. 458-467, 1971) . The dura mater was homogenized twice in phosphate-buffered saline (PBS, pH 7.4), extracted for 24 hours in 2:1 chloroform/methanol, and acid-hydrolyzed under reflux and nitrogen for 36 hours in 6 liters of 6 N HC1. The acid was evaporated at 40°C The residue was dissolved in water and pH adjusted to 7.4 (NaOH) . The material was applied to a 5 x 150 cm column of Bio-Gel P-2 fine (Bio-Rad Laboratories, Rockville Centre, NY) equilibrated with 0.02 M Hepes (pH 7.4) containing 0.15 M NaCl. Fifteen milliliter fractions were collected at a flow rate of 1 ml/min. Fractions containing the 335/385 fluorophore were pooled, adjusted to pH 8.5 with NaOH and rotary evaporated. The fluorophore was extracted with methanol to remove some

SUBSTITUTESHEET(RULE25)

of the salts. The methanol, in turn, was evaporated and the residue dissolved in 10 ml of water and acidified with concentrated HCl. Purification was achieved by multiple injections/peak collections using reverse-phase C-18 HPLC and a water/acetonitrile solvent containing consecutively trifluoroacetic acid (TFA) , n-heptafluorobutyric acid (HFBA) , and again TFA as counterions. The final product was judged pure by virtue of a single ninhydrin positive spot on paper chromatography and a single UV and fluorescent HPLC peak under various chromatographic conditions using a reverse-phase column.

Synthesis and Purification of Pentosidine from a Synthetic System

Three liters containing 100 mM each of L-arginine, L- lysine and D-ribose, at pH 7.3, were heated for 1 hr at 80°C The cooled mixture was poured onto a Buchner funnel filled with Dowex 50 x 4 400 ion-exchange resin (Aldrich Chem. Co. , Inc. , Milwaukee, WI) equilibrated according to conditions of Boas (Boas, N.F., J. Biol. Chem.. 204, pp. 553-563, 1953) . The resin was washed with 2 liters each of water and 1 M pyridine, followed by elution of pentosidine- containing material with 1 liter of 2 N NaOH. The material was then adjusted to pH 7.4 with HCl, concentrated by rotary evaporation, passed through a Bio-Gel P-2 column and purified by HPLC as described above. The material was also

SUBSTITUTESHEET(RULE25)

chromatographed on Whatman 17 Chr paper (Whatman Inc. , Clifton, NJ) using 1:1 water/pyridine. Upon elution from the paper, the material was reinjected and collected by reversed-phase HPLC using TFA as the counterion.

High Performance Liquid Chromatography (HPLC)

A Waters HPLC (Waters Chrom. Div. , Milford, MA) with Model 510 pumps, U6K injector, and a 680 controller was used. The effluent was monitored at 385 nm (excitation at 335 nm) with a J4-8202 Aminco-Bowman spectrophotofluorometer (SLM Instru. , Inc., Urbana, IL) equipped with a 9 μl continuous flow cell. Separations were made on either a 4.6 mm (analytical) or 1.0 cm (semi- preparative) x 25 cm Vydac 218TP (10 micron) C-18 column (The Separations Group, Hesperica, CA) by application of a linear gradient system of 0-17% acetonitrile from 10 to 97 min, . with either TFA or HFBA as counterion at a flow rate of l (analytical) or 2 ml/min (semi-preparative) . Pyridoxamine (Sigma Chem. Co., St. Louis, MO) which autofluoresces was used as an internal standard.

For analytical purpose, the HPLC program was shortened such that a linear gradient of 10-17% acetonitrile was applied from 0-35 min. with HFBA as the counterion (see Figure 5) .

SUBSTITUTESHEET(RULE25)

Incubation of Sugars with L-lysine and L-arginine In Vitro Sugars were incubated with L-lysine and L-arginine at 80 ^β C for 1 hour in PBS. Each sugar and amino acid were present at concentrations of 100 mM in a total volume of 2 ml in 13 x 100 mm test tubes placed in a Reacti-Therm heating block (Pierce Chem. Co., Rockford, IL) .

In Vitro Incubation of Young Collagen with Pentoses

A pool of young dura mater (average age 15 years) obtained at autopsy was homogenized in PBS and extracted for 24 hours in 2:1 chloroform/methanol. Dry blotted tissue (0.5 g wet weight) was incubated with 100 mM each of the pentose sugars, L-arginine and L-lysine in 12 ml of PBS containing 10 μl each of toluene and chloroform. After incubation at 37°C for 6 days, 0.3 g wet tissue weight was withdrawn from each tube and washed three times with 5 ml portions of PBS and water, respectively. Samples were acid-hydrolyzed in 2 ml of 6 N HCl for 24 hours. Following evaporation of the acid, the material was reconstituted in 1 ml of water. Hydroxyproline content was determined as described (Hamlin, C.R. , Luschin, J.H., and Kohn, R.R., EXP. Gerontol. 13, pp. 415-523, 1978) and equalized among samples.

SUBSTITUTESHEET(RULE25)

Determination of Pentosidine in Tissues

Tracheal cartilage, cortical bone (iliac crest), aorta, kidney, cardiac muscle, lung, liver, skin, dura mater and lens were obtained at autopsy from elderly subjects. Cartilage, bone and aorta were decalcified. All tissues were minced and extracted with 4-5 changes of PBS before lyophilization. Red blood cells, obtained by centrifugation of human blood, were washed three times in PBS and lyophilized. The following were gifts: purified isolated human renal glomerular basement membrane from Dr. Edward C Carlson (University of North Dakota) ; human and rat-cultured glomerular mesangial cells from Dr. John R. Sedor (Case Western Reserve University, School of Medicine) ; and mixed human fibroblasts cultured for 14 days on rat tail tendon collagen-coated petri dishes from Dr. Irwin A. Schafer (Case Western Reserve University, Cleveland Metropolitan General Hospital) . Human placental, bovine Achilles tendon and calf skin collagens were purchased from Sigma Chem. Co. (St. Louis, Mo) . Approximately 15 mg of each sample were acid- hydrolyzed in 2 ml of 6 N HCl for 24 hours. The acid was evaporated and pentosidine was quantitated by HPLC after reconstituting samples with water.

SUBSTITUTESHEET(RULE25)

Spectroscopy

Absorption spectra were recorded with a Hewlett- Packard (HP) 8452A diode array spectrophotometer connected to an IBM PC/AT computer (Hewlett-Packard, Inc., Avondale, PA; IBM Corp., Boca Raton,FL). Fluorescence spectra were recorded with a J4-8202 Aminco-Bowman spectrophotofluorometer (SLM. Instru. , Inc., Urbana, IL) .

Samples for proton NMR spectroscopy were exchanged three times with deuterium oxide (D ₂ 0) under a nitrogen atmosphere. The sample contained in 400 μl of 100% D _; 0 was transferred to a 5mm NMR tube and scanned for 10 min. in a

400 MHz spectrometer (MSL 400, Bruker Instru., Inc.,

Billerica, MA) . The following conditions were used (Figure

3): spectrometer frequency, 400.13 MHz; spectral width, 1 ppm — 400.13 Hz; Hz/point = 0.244; acquisition time, 2.048 s (native), 1.024. s (synthetic); number of scans, 944

(native), 400 (synthetic); temperature, 297°K; recycle delay, 5 s; pulse width, 6.65 μs corresponding to 90°.

TPS(3-(trimethylsilyl)-1-propanesulfonic acid) was used as an internal standard. For twodimensional H,H-correlated

(COSY) spectroscopy, the sample was scanned overnight.

Mass-spectrometry analyses were performed by Dr.

Douglas Gage at the NIH mass-spectrometry facility in the

Department of Biochemistry, Michigan State University, East Lansing, MI. Molecular weights were determined by fast atom bombardment (FAB) spectroscopy with a JEOL HX 110HF

SUBSTITUTESHEET(RULE25)

double focusing mass spectrometer. Analysis was initially conducted at low resolution

(1000) at accelerating voltage of 10KV. Samples were dissolved in 0.1% TFA and mixed with an equal volume of glycerol. Ions were formed by FAB with a 6 KeV beam of Xe atoms. Spectra which were generated by FAB CAD MS/MS analysis (collisionally activated dissociation tandem mass spectrometry) made use of a JEOL DA-5000 data-system- generated linked scans at constant B/E. Helium was used as the collision gas in a cell located in the first field-free region and the pressure was adjusted to reduce the abundance of the parent ion by 75%. FAB high resolution mass analysis was performed at resolution 20,000 by peaking matching on the glycerol matrix ion at M/Z 369.

RESULTS

Isolation and Purification of Pentosidine from Tissue.

A pool of dura mater (600 g wet weight) obtained at autopsy from elderly donors (average age 77 years) was acid-hydrolyzed and fractionated by Bio-Gel P-2 gel filtration chromatography (see Figure 1) . The bulk of the fluorescence material eluting together with salt was pooled, dried by evaporation and extracted into methanol. After evaporation and reconstituting in water, the fluorescent material was

SUBSTITUTESHEET(RULE25)

purified to homogeneity by repeated injections on reverse phase HPLC. Total fluorophore recovered was 1 mg.

Structure Elucidation of Pentosidine.

The fluorophore was characterized by absorption, fluorescence, Η-NMR, and mass spectroscopical properties. Its UV and fluorescence maxima (Figure 2, top) were identical with those of the previously described unhydrolyzed fluorophore (Sell, D.R. , and Monnier, V.M. , Conn. Tiss. Res. 19, pp. 77-92, 1989) suggesting that no damage occurred as a consequence of hydrolysis. A peculiarity was noted in fluorescence-excitation intensities which varied with pH, being highest at pH 2 and 12, and completely quenched at pH 9 (see Figure 2).

Crucial structural information was obtained from the Η-NMR spectrum (see Figure 3, top; also Figure 4) which showed two doublets, a and b at 7.78 and 7.94 ppm, respectively, that were coupled with the triplet c at 7.22 ppm (Figure 4) as revealed by a COSY experiment (not shown) . This configuration suggested the presence of three aromatic protons in a pyridinium molecule with substitutions in positions 5 and 6. The two uncoupled triplets at 3.9 and 3.95 ppm suggested the presence of two α-protons compatible with the presence of two amino acids. Two other triplets (d and e in Figure 3) were observed at 3.6 and 4.6, both coupled with aliphatic protons at 2.0

SUBSTITUTE SHEET (RULE 25)

ppm. By comparison with published spectra of pyridinoline (Fujimoto, D., Moriguchi, T. , Ishida, T. , and Hayashi, H., Biochem. Biophvs. Res. Commun. 84, pp. 52-57, 1978) (Deyl, Z., Macek, K. , Adam, M. , and Vancikova, Biochem. Biophys. Acta, 625, pp. 248-254, 1980) (Ogawa, T. , Tsuda, T.O.M., and Kawonishi, Y. , Biochem. Biophvs. Res. Commun.. 107, pp. 1252-1257, 1982), lysine emerged as a likely component of the fluorescent molecule.

FAB high resolution mass spectrometry showed a M/Z of 379.2069 compatible with the empirical formula C ₁₇ H ₂₇ N ₆ 0 ₄ . Taken together, the data suggested the possible presence of an imidazo [4,5b] pyridinium ring comprising a five-carbon moiety (highlighted as bold lines in the structure below) with a lysine and arginine side chain. This configuration suggested that crosslinking of the two amino acids might have occurred as a consequence of Maillard reaction with a pentose.

In Vitro Synthesis of Pentosidine In order to confirm the pentose-derived nature of the native fluorophore 100 mM each of L-lysine, L-arginine and

D-ribose were heated for 1 hour at 80°C Injection of a small amount of this synthetic material on HPLC revealed a major fluorescent peak co-eluting with the native fluorophore.

SUBSTITUTE SHEET(RULE25)

(see Figure 5) In order to substantiate the proposed structure, the synthetic fluorophore was prepared preparatively and purified. Total yield was 21 mg; i.e.,

0.02% of the reactants.

Spectroscopical Comparison of Native and Synthetic Pentosidine

The synthetic fluorophore showed the same UV and fluorescence spectra, including pH effects, as those of the native fluorophore (Figure 2, bottom). The molar absorption coefficients of the native and synthetic compounds were determined to be 4522 and 4195 in 0.1 N HCl, respectively.

Η-NMR spectra of the synthetic and native fluorophores were similar except for a small shift of the α-proton triplets at 4.08 and 4.16 ppm attributable to a pH or concentration effect (Figure 3) . Other structural assignments are shown in Figure 3.

Fragmentation patterns (Figure 6) of synthetic and native compounds obtained by FAB CAD MS/MS analysis were also identical except for minor differences in peak intensities attributed to differences in operating conditions of the instrument since analyses were made 6 months apart. FAB high resolution analyses showed measured M/Z of 379.2069 and 379.2091 for the native and

SUBSTITUTE5HEET(RULE25)

synthesized fluorophores, respectively (± 0.4 ppm instrument error) . The calculated mass of the proposed compound is 379.4392.

Origin of the Imidazo T4,5bl Pyridinium Ring The spectroscopical data from the native and synthetic fluorescent molecules leaves little doubt as to the nature and structure of the newly discovered crosslink. The complete aromatization of ribose in the formation of the pyridinium ring, however, suggests that the iso ers of ribose; i.e., arabinose, xylose and lyxose, can also mediate the same reaction. To test for this possibility, and to investigate the structure requirements of reducing sugars for the formation of the fluorescent molecule, hexoses and pentoses were reacted with equimolar amounts of Llysine and L-arginine at 80°C for 1 hour. The results are set forth below in Table I.

SUBSTITUTESHEET(RULE25)

TABLE I

The Effects of Incubation of Various Sugars with

Lysine and/or Arginine for l Hour at 80°C on the

Formation of Pentosidine

(Quantitated according to the conditions of Figure 5)

(+) added, (-) not added, (ND) not detected.

The results in Table 1 indicate that all three aldopentoses (xylose, arabinose and lyxose) could serve as

SUBSTITUTESHEET(RULE25)

precursors of the imidazo [4,5b] pyridinium ring. This observation led the present inventors to name the fluorophore "pentosidine". None of the hexoses tested, however, were able to generate this compound. Extremely low levels were detected with 2-deoxy-Dribose, ribose-5- phosphate and pentuloses suggesting that commercial preparations of these sugars contain small amounts of pentosidine precursors. Pentosidine could also be synthesized by direct incubation at physiological pH and temperature of young collagen with pentoses (Table II) . The highest yield was obtained with D-ribose. The addition of free lysine or arginine blocked impart the synthesis of pentosidine presumably by trapping of free ribose or intermediates of the Maillard reaction that might be involved in pentosidine synthesis (Table II, Experiment 1) . A possible biosynthetic mechanism for pentosidine formation is depicted in Figure 7.

SUBSTITUTESHEET(RULE25)

TABLE II

The Effects of Incubation of Young Collagen With Pentose Sugars, Lysine, and/or Arginine for 6 Days at 37°c on the Formation of Pentosidine ⁰

(Quantitated according to the conditions of Figure 5)

Pentose L-Lysine L-Arginine Pentosidine (100 mM) (100 mM) (100 mM) (pmole/mg Collagen)

Experiment 1

57 326 131 118 107

69

None 46

D-Xylose 125

D-Arabinose 109

D-Ribose 288

D-Lyxose 168

(-) not added, (+) added

Pentosidine in Various Biological Specimens

The presence of pentosidine was studied in a variety of tissues by HPLC. Quantitation in aging human dura mater

SUBSTITUTESHEET(RULE25)

revealed a linear 10-fold increase throughout life which reached approximately 250 pmol/mg collagen in late life (see Figure 8) . In a separate study on pentosidine level in human skin, the progression was exponential in late life but reached only 75 pmol/mg collagen suggesting a higher turnover of skin than dura mater (See Figure 9) . Pentosidine was also detected in crude preparations of human heart, aorta, lungs, cartilage, bone, tendon, liver, renal cortex and medulla, and a pure preparation of glomerular basement membrane obtained after proteolytic digestion (see Table III below) . The question of whether pentosidine in these tissues originated primarily from cellular or extracellular matrix was not investigated at this point. However, pentosidine was also detected in red blood cell and plasma proteins (Table III and Figure 12) suggesting that the ability of pentoses to crosslink proteins is not limited to the extracellular matrix. No pentosidine was detected in commercial preparations of Type I, III, IV and V soluble collagens, but a small level was detected in a commercial preparation of insoluble Type I collagen obtained from bovine tendon (Table III) .

SUBSTITUTESHEET(RULE25)

TABLE III Summary of Pentosidine Levels in Different Tissues

(Quantitated according to the conditions of Figure 5) mole Tissue

HUMAN:

Dura Mater

Skin

Tracheal Cartilage

Cortical Bone

Aorta

Cardiac Muscle

Lung

Liver

Kidney Cortex

Kidney Medulla

Purified Isolated Glomerular Basement Membrane

Red Blood Cells

Blood Proteins

Lens

Placenta (Commercial Types III, IV, V Soluble Collagens) ND' ND

SUBSTITUTE SHEET (RULE 25)

OTHER :

Calf Skin (Commercial Type I

Soluble Collagen) ND ND

Bovine Tendon (Commercial Type

I Insoluble Collagen) 10 7

CELL CULTURE:

Human Fibroblasts ^b 345 —

Collagen Matrix (Blank) ⁶ 25 —

* - ND, not detected ^b - Mixed fibroblasts cultured for 14 days on rat tail tendon collagen-coated petri dishes ^c - Represents a control consisting of collagen-coated petri dishes containing medium incubated for 14 days without cells.

A very low level was detected in the human ocular lens, a tissue with a metabolism significantly different from that of tissues rich in nucleated cells. Finally, and unexpectedly, pentosidine was detected in human fibroblasts grown in culture (Table III) . A high quantity was also detected in cultures of human glomerular mesangial cells grown in a pentose-free medium. However, additional studies will be needed to determine its origin.

The effect of age alone on pentosidine level in skin biopsies taken from individuals without diabetes or renal failure is shown in Figure 9. The data used to develop this graph was taken from a testing of human skin samples. Normality and variance equalities were tested according to the procedures of Shapiro and Wilke, Shapiro SS and MB

SUBSTITUTESHEET(RULE25)

Wilke, (©1965) , "An Analysis of Variance Test For Normality", Biometrika. 52: 591-611, and Steele and Torrie, Steele et al., "Principles and Procedures of Statistics", McGraw-Hill. Inc.. New York, 471, 542-543 (©1980). Data were transformed by use of the square root transformation according to Steele and Torrie. Simple and multiple regression techniques were from Neter and Wasserman with use of methods available in the SPSS/PC Plus Statistical Software. Neter et al., "Applied Linear Statistical Models", Richard D. Irwin. Inc.. 21-393 (©1974). Confidence interval for the regression line was computed using the error of prediction formula given by Armitage. Armitage et al. , "Statistical Methods in Medical Research", John Wylie & Sons. 163-165 (©1971) . The tissue samples from which this data was collected were assayed for pentosidine level as described hereinabove. Figure 9 shows a curvilinear, exponential increase with age that corresponded to a six-fold increase between 10 and 100 years. It should be noted that beyond the age of 80 years pentosidine levels started to spread. Taking this into account, the regression line shown in Figure 9 was computed using data points from ages 8 to 80.

Figure 10 demonstrates the relationship between age, diabetes, renal disease, and pentosidine level. The data was collected from the same pool of skin biopsies used to develop the standard curve shown in Figure 9. In this

SUBSTITUTESHEET(RULE25)

Figure 10, levels are expressed relative to a normal range computed as a 95% confidence interval determined and reproduced from the regression line of Figure 9. Subjects without renal failure are represented by open symbols, while subjects with renal failure are represented by closed symbols. A closed square represents a non-diabetic with renal failure; an open triangle represents a type I diabetic without renal failure; a closed triangle represents a type I diabetic with renal failure; an open circle represents a type II diabetic without renal failure; and, a closed circle represents a type II diabetic with renal failure. This figure clearly suggests the presence of a profound abnormality of pentose metabolism in subjects with uremia. The abnormality would appear to be accentuated by diabetes and also to be present to a milder degree in diabetic subjects that are not uremic.

Pentosidine has also been detected in human red blood cells and plasma proteins proving that the abnormalities described above in relation to Figure 10 can be assessed by using a blood-based, instead of a skin-biopsy, assay for pentosidine. Figure 10 demonstrates the distribution of pentosidine values in acid hydrolysate of plasma and hemolysate for control (CON) subjects, diabetic (DB) subjects, and uremic (ESRD) subjects whose plasma was obtained and assayed by combined reverse-phase, high- performance liquid, and ion-exchange chromatography. The

SUBSTITUTESHEET(RULE25)

results show that mean plasma protein pentosidine levels are elevated 2.5-fold in diabetic subjects and 23-fold in uremic subjects.

In addition, detection of the pentosidine compound may also aid in determining who among diabetic subjects is at risk of developing diabetic complications. Figure 12 sets forth the results of testing done to confirm the relationship between pentosidine level and severity of complications. In this figure, the level of pentosidine is presented as a function of the cumulative grade of all complications classified into one of three categories. The ranges for which were 0-2, 3-5, and 6-8. In this case, the increase in pentosidine levels with cumulative grade of complications was highly significant. Mean levels for subjects suffering from the most severe combined complications (Category 6-8) was elevated 2.6-fold, whereas it was elevated only 2-fold in subjects with mild or no complications (Category 0-2) compared with control subjects. The data used to compile Figure 12 is set forth in Table IV summarizing the skin pentosidine level in relation to type and severity of diabetic complication.

SUBSTITUTESHEET(RULE25)

TABLE IV

Summary of skin pentosidine in relation to type and severity of diabetic complications.

Pentosidine ± SD Group and Complication n fpmol/mg collagen)*

20.6 ± 3

42.7 ± 7

44.4 ± 12 50.1 ± 9

45.3 ± 11 43.0 ± 7 50.7 ± 5

45.0 ± 10 43.9 ± 11 51.0 ± 9

40.7 ± 6

46.5 ± 11 49.4 ± 10

♦Values were age-adjusted to 35 to make comparisons among groups.

In Table IV, grades 0, l, and 2 represent the severity of the complication. Retinopathy was assessed by an opthomologist using funduscopy and was graded as no fundus abnormality (0) , background retinopathy (1) , or proliferative retinopathy (2) . In similar manner, grades 0, l, and 2 represent the severity of nephropathy, determined with respect to urinary protein output; arterial stiffness, assessed by an index of aortic pulse-wave velocity; and, joint mobility, graded on the basis of hether mobility was limited to fingers only or fingers and other large joints of the subject.

SUBSTITUTESHEET(RULE25)

Discussion The discovery of an age-related accumulation of pentosidine in human extracellular matrix is the first molecular evidence for the involvement of reducing sugars in protein crosslinking. In preliminary studies, Kohn, et al. (Kohn, R.R. , Cerami, A., and Monnier, V. M. , Diabetes 33, pp. 57-59, 1984) demonstrated that rat tail tendons incubated with reducing sugars became rapidly crosslinked. The crosslinking rate as measured by tail tendon breaking time in urea was much greater for ribose than glucose and was accompanied by formation of collagen-linked fluorescence. While the study which resulted in the present invention was in progress, Tanaka, et al. (Tanaka, S., Avigad, G. , Eikenberry, E.F., and Brodsky, B., J. Biol. Chem.. 263 pp. 17650-17657, 1988) reported the isolation of highly fluorescent dimers of α chains crosslinked in triple-helical regions of ribose-incubated rat tail tendon collagen.

Although these studies suggested that the pentose could mediate crosslinking in vitro. the discovery of pentose-mediated crosslinking in vivo is quite unexpected and raises a number of biochemical and biological questions concerning the origin and role of pentosidine. The absence of detectable pentosidine in solutions of glucose incubated for l hour at 80 ^β C with equimolar lysine and arginine suggests that glucose and its Amadori product are unlikely

SUBSTITUTESHEET(RULE25)

precursors of pentosidine. However, little browning was yet detectable after 60 min. and a fragmentation of glucose or its Amadori product into a pentose upon prolonged reaction is not excluded. Similarly, studies will be needed to evaluate the possible contribution of glyco- conjugates to pentosidine recovered from acid-hydrolyzed biological specimens. In this sense, the data presented in Table III should be considered as preliminary. Albeit this word of caution, there is little doubt that pentosidine forms spontaneously in aging since it increased in an enzymatic hydrolysate of human dura mater with age (Sell, D.R. , and Monnier, V.M. , Conn. Tiss. Res. 19, pp. 77-92, 1989).

In contrast to the extensive literature on blood and tissue levels of glucose, however, only a scant amount of information is available on the source and the level of free pentoses in tissue and body fluids. All three sugars, ribose, xylose and arabinose, have been detected in the urine with excretion rates of 5, 8.5 and 14 μg/min, respectively (Bell, D.J., and Talukder, M. , Q-K, Clin. Chim. Acta. 40, pp. 13-20, 1972). Total pentose level in human plasma has been estimated at 44 μM (McKay, E. , Clin. Chim. Acta. 10, pp. 320-329, 1964), but no information is available on the concentration of particular pentoses. A number of observations suggest that ribose or one of its metabolites is a likely precursor of pentosidine in

SUBSTITUTESHEET(RULE25)

vivo. First, of all tested pentoses, ribose was the most reactive sugar in the synthesis of pentosidine (Table II) . This observation is in agreement with previous i determinations of the chemical reactivity of ribose (Overend, W.G. , Peacoke, A.R. , and Smith, J.B., J. Chem. Soc.. pp. 3487-3492, 1961). Second, free arabinose and xylose are thought to arise primarily from alimentary sources, mainly through the ingestion of fruits and the bacterial degradation of xylans in the intestine (McKay, E. , Clin. Chim. Acta. 10, pp. 320-329, 1964) (Date, J.W. , Scand. J. Clin. Lab. Invest.. 10, pp. 155-162, 1958). Thus, it is unlikely that these sugars would explain pentosidine formation in cell culture. Third, lyxose has been only associated with heart muscle (Pailares, E.S., and Garza, H.M. , Arch. Biochem.. 22, pp. 63-65, 1949). Finally, the detection of high pentosidine levels in cell culture (Table III) is strongly suggestive for leakage or release of significant amounts of free ribose or its metabolites as a consequence of accelerated ribonucleotide turnover, cellular turnover or cell death. In this regard, a possible source for a precursor of pentosidine could come from ADP-ribosylation reactions which play a crucial role in many cellular functions, including DNA repair mechanisms which are thought to play a role in cellular aging (Ueda, K., and Hayaisha, 0., Ann. Rev. Biochem. 54, pp. 73-100,

SUBSTITUTESHEET(RULE25)

1985) . Yet, these propositions are speculations that will need to be addressed experimentally.

Pentosidine is of significance for gerontological research for two reasons. First, pentosidine may contribute to the age-related stiffening of tissues by crosslinking of the extracellular matrix. The extent of crosslinking can be estimated by the relative amount of crosslinks found in old human dura mater collagen assuming a molecular weight of 300,000 for the triple-helical region. The presence of 250 pmol/mg collagen translates into 7.5% modification (0.075 mol/mol of collagen). This would be compatible with a 2 to 3-fold decrease in collagen digestibility according to the estimate by Vater, et al. (Vater, CA. , Harris, E.D., and Siegel, R.C., Biochem. J.. 1881, pp. 639-645, 1979). It is also possible that additional pentose derived crosslinks would form during aging because pentosidine was only one of many compounds present in the reaction mixture of arginine, lysine and ribose. Second, pentosidine may serve as a molecular marker of the aging process and its availability should greatly facilitate studies on longevity and the potential role of Maillard-mediated damage by pentoses in the life- span limiting process.

The invention has been described with reference to the preferred embodiment. Obviously, modifications and alterations will occur to others upon reading and

SUBSTITUTESHEET(RULE25)

understanding the preceding detailed description. It is intended that the invention be construed as including all such alterations and modifications insofar as they come within the scope of the appended claims or the equivalents thereof.

Example 2 The age-related increase of pentosidine proceeds at different rates, depending on the type of tissue (dura mater, skin, or human glomerular basement membrane) . This data suggest a relation between the advanced Maillard reaction and mechanisms of aging and diabetic compli¬ cations. Also, pentosidine formation from glucose is thought to be related to oxidative stress, in that oxygen is required for its synthesis. Pentosidine is also present in plasma proteins. Thus, determination of plasma pentosidine concentrations provides information on the extent of damage to tissues, although it may represent only the tip of the iceberg of modifications by the Maillard reaction. Until recently, pentosidine has been quantified by HPLC after acid hydrolysis of samples. Accurate determina¬ tion of low concentrations could be obtained only with a cumbersome double-chromatographic system. For this study, the inventors raised antibodies against pentosidine and developed a simple ELISA system to determine pentosidine

SUBSTITUTESHEET(RULE25)

concentrations in plasma and tissue protein. A highly specific antibody was obtained that allowed the inventors to obtain reproducible data that compared well with those obtained by HPLC.

Materials and Methods

Preparation of Reagents

Immunoqen. Synthetic pentosidine was prepared along the lines mentioned above. Solubilization of keyhole limpet hemocyanin (KLH) and conjugation of pentosidine to KLH was done by using the method of Staros et al. (Staros, J.V. , Wright, R.W. , Swingle, D.M. Enhancement by N- hydroxysulfosuccinimide of water-soluble carbodiimide- medicated coupling reaction. Anal Biochem.. 156, pp. 220-2 (1986)). Briefly, 7 mg of KLH (Calbiochem, San Diego, CA) was incubated with 15 μmol of pentosidine and 0.1 mmol of l-ethyl-3-(3-dimethylaminopropyl) carbodiimide-HCI (Pierce Chemical Co., Rockford, IL) in the presence of 1.4 mg of N-hydroxysulfosuccinimide (Pierce) at room temperature for 4h and at 4°c overnight. Pentosidine incorporation into KLH was 155 μmol/g protein, as determined by HPLC (Odetti, P., Fogarty, J. , Sell, D.R. , Monnier, V.M. Chromatographic quantitation of plasma and erythrocyte pentosidine in diabetic and uremic subjects. Diabetes. 41, pp. 153-9 (1992)) .

SUBSTITUTESHEET(RULE25)

Antibodv. KLH-pentosidine (250 mg) emulsified with an equal volume of complete Freund ^, s adjuvant was infected into two New Zealand White rabbits intradermally and intramuscularly. At 2- to 3-week intervals, KLH-pentosidine emulsified with an equal volume of incomplete adjuvant was injected as a booster. Sera were obtained 7-10 days after each injection and were titrated for antibody. After the titer reached a plateau, the rabbits were killed by heart puncture, and the IgG from their blood was purified by chromatography on Protein G-

Superose (Pharmacia, Uppsala, Sweden) .

Bovine serum albumin fBSA)-ribose. BSA (essentially fatty-acid-free; Sigma Chemcial Co., St. Louis, MO), 50 g/L, was incubated with 0.1 mol/L ribose in phosphate- buffered saline (PBS), pH 7.4, at 37 ^β C Aliquots were withdrawn on days 0, 1, 3, 5, 7, 14, 30, 60 and 90 and dialyzed against PBS.

Lvsozvme-ribose. Lysozyme (from chicken egg white;

Sigma), 20 g/L, was incubated with 0.5 mol/L ribose in 0.2 mol/L sodium phosphate (pH 7.4) at 37°C Aliquots were withdrawn on days 0, 1, 4, 8, 11, and 16 and dialyzed against PBS.

SUBSTITUTE SHEET (RULE 25)

Sample Pretreatment

Enzv e digestion. Plasma and BSA-ribose were incubated with 20 g/kg Pronase E (EC 3.4.24.31; Type XXV, Sigma) at 37°c for 24 h and filtered by centrifugation (Millipore, Bedford, MA) at 5000g for 30 min to remove the enzyme. The digestions were judged by ninhydrin assay (Moore, s., Stein, W.H. A modified ninhydrin reagent for the photometric determination of amino acid and related compounds. J. Biol. Chem.. 211, pp. 907-13 (1954)). Plasmas were collected from nondiabetics, diabetics, and patients with end-stage renal disease by the Laboratory of Clinical Chemistry at the University Hospitals of Cleveland. Plasmas were partially delipidated by centrifugation at 100,000g for 30 min. and removal of the top layer.

Processing skin collagen. Human skin samples were obtained from tissue acquired by the inventors laboratory at autopsy. Insoluble collagen was prepared from skin samples as previously described (Monnier, V.M. , Vishwanath, V., Frank, K.E., Elmets, CA. , Dauchot, P., Kohn, R.R. Relation between complications of type I diabetes mellitus and collagen-linked fluoresceince. N. Engl. J. Med.. 314, pp. 403-8 (1986)). After removal of the subcutaneous fat and epidermis with a sharp razor blade, the tissue was homogenized in PBS, extracted with chloroform-methanol, and lyophilized. These extracts, redissolved in 6 mol/L

SUBSTITUTESHEET(RULE25)

hydrochloric acid (1-2 g of collagen per liter) , were treated with a nitrogen stream for 1 min and then heated in an oven at 110°C 16 h. Subsequently, samples were concentrated with a Speed Vac concentrator equipped with an HCl trap (Savant, Farmingdale, NY) , filtered with a 4-μm (pore size) filter, and freeze-dried.

For ELISA, hydrolyzed samples were reconstituted in enough 0.5 mol/L NaOH and 0.2 mol/L sodium phosphate buffer to adjust to neutral pH and diluted with PBS to the desired concentration. For HPLC, hydrolyzed samples were reconstituted in 1 mL of water containing 10 mmol/L heptafluorobutyric acid. Collagen content was determined according to the method of Stegemann and stalder (Stagemann, H. , Stalder, K. Determination of hydroxyproline. Clin. Chim. Acts. 18, pp. 267-73 (1967), assuming a content of 140 g/kg hydroxyproline (Hamlin, C.R. , Kohn, R.R. Evidence for progressive age-related strucutral changes in post mature human collagen. Biochim. Biophvs. Acta.. 236, pp. 458-67 (1971)). Plasma and BSA- ribose were also hydrolyzed by the same methods.

Assay Procedures

ELISA. Assay of pentosidine and cross-reactivity testing with various compounds were done by competitive

ELISA, as previously described in the inventors' laboratory (Miyata S, Monnier VM. Immunocytochemical detection of

SUBSTITUTESHEET(RULE25)

advanced glycation end product in diabetic tissues using monoclonal antibody to pyrraline. J. Clin. Invest. 1992; 89:1102-12.). The coating agent was BSA conjugated to pentosidine, the amount of pentosidine incorporated into BSA being 500 μmol/g BSA as estimated by HPLC. The BSA- pentosidine conjugate (50 μg/L) was dispensed into each well of a microtiter plate and incubated for 1 h at 37 ^β C After the plate was washed with PBS, the wells were blocked with 200 μL of 10 g/L ovalbumin (Grade V; Sigma) for 1 h at 37°C and then washed with PBS containing 0.5 mL/L Tween 20. Antibody, 100 μL (25 ng of IgG) per well, was preincubated with pentosidine or various compounds (Fig. 13) at 37 ^β C for 3 h and then applied to the plate, which was incubated with shaking at room temperature for 1 h. After washing, 100 μL of l000-fold-diluted alkaline phosphate-linked goat antibody to rabbit IgG (Boehringer Mannheim, Mannheim, Germany) was added, and the samples were incubated at room temperature for 1 hr. After washing, p-nitrophenyl phosphate substrate solution (50 μg/L) was added, and the samples were incubated at room temperature until adequate color developed. The enzyme reaction was stopped by adding 50 μL of 2.5 mol/L NaOH to each well. Absorbance at 405 nm was measured with a Microplate Reader (Bio-Rad, Richmond, CA) , with 655 nm as reference wavelength. Pentosidine was quantified from a calibration curve that ranged from 10" ³ to 10 ³ pmol/well.

SUBSTITUTESHEET(RULE25)

HPLC For comparison, the inventors determined pentosidine in acid-hydrolyzed samples by reversed-phase HPLC. A solution of hydrolysate was injected into a HPLC system (Waters Millipore, Milford, MA) consisting of a Model 712 WISP autosampler, two Model 510 pumps, a Model 680 gradient controller. Separations were made with a 25 x 0.46 cm C ₁₈ Vydac Type 218P (10 mm) column (Separations Group, Hesperia, CA) with a linear gradient program of 100- 170 mL/L acetonitrile (Fisher Scientific, Fair Lawn, JH) from 0 to 35 min in mobile phase containing water and 1 mL/L heptafluorobutyric acid as a counterion. The effluent was monitored with a Model 470 fluorescence detector (Waters) set at excitation/emission wavelengths of 335/385 nm. Attenuation was 4, gain x 100, and filter at 1.5 s. The material eluting in the pentosidine peak was collected, evaporated, and rechromatographed on a cation-exchange column (SP-5PW; Waters) eluted with a curvilinear gradient of NaCl (0-0.6 mol/L) from 0 to 40 min in 0.02 mol/L sodium acetate buffer (pH 4.7). This procedure was used especially for low concentrations of pentosidine (i.e., <20 nmol/g protein) (Odetti, P, Fogarty, J, Sell, D.R. , Monnier, V.M. Chromatographic quantitation of plasma and erythrocyte pentosidine in diabetic and uremic subjects. Diabetes, 41, pp. 153-9 (1992)). Tiηmunoblotting. The lysozyme-ribose reaction mixture (5 μg of protein each) was separated by sodium dodecyl

SUBSTITUTESHEET(RULE25)

sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli, U.K. Cleavage of structural proteins during the assembly of th ehead of bacteriophage T4. Nature. 227, pp. 680-5 (1990)) at 200 V (constant) for 1 h with a Protein Miniblot (Bio-Rad) and a 12% acrylamide gel. After transfer to an Immobilon (Millipore) membrane for 1 h at 110 V (constant) , the membrane was incubated with 30 g/L BSA for 1 h, followed by two 5-min washes with Tris-buffered saline (TBS) containing 0.5 mL/L Tween 20 (TTBS) . The membrane was then incubated with antibody to pentosidine (0.5 mg/L IgG) in 10 g/L BSA in TTBS. For development, the membrane was washed twice (5 min each) in TTBS, then incubated with 1000-fold diluted goat anti-rabbit IgG coupled to alkaline phosphatase (Boehringer Mannheim) in 30 g/L lysozyme in TTBS for 16 min. The membrane was again washed twice with TTBS, once with TBS, and incubated with 5-bromo-4-chloro-3-indolyl phosphate (0.15 g/L) and nitroblue tetrazolium (0.30 g/L) (mixed tablet; Sigma) solution until specific bands were clearly visible. The color development was stopped by washing the membrane in distilled water. To assess inhibition of the test by free pentosidine, the inventors preincubated antibody to pentosidine with 50 nmol of free pentosidine at 37°C for 3 h.

SUBSTITUTESHEET(RULE25)

RESULTS Titer of anti-pentosidine antibody. The titer of pentosidine antiserum, detected by its binding to BSA-pen- tosidine coating agent, increased as a function of time and reached a plateau at 7 weeks after the initial immunization injection (Fig. 14) . Serum from the rabbit showing the highest titer at 13 weeks was subjected to Superose-Protein G chromatography to isolate a pure IgG fraction.

ELISA. The inventors used ovalbumin as a blocking agent because it has only trace amounts of pentosidine (Odetti, P., Fogarty, J. , Sell, D.R. , Monnier, V.M. Chromatographic quantitation of plasma and erythrocyte pentosidine in diabetic and uremic subjects. Diabetes. 41, pp. 153-9 (1992)). Careful optimization of the conditions for competitive ELISA showed best sensitivity with 5 ng of coating agent and 25 ng of anti-pentosidine IgG per well. The reaction of the antibody and BSA-pentosidine in the competitive ELISA system was inhibited by free pentosidine. The detection limit, calculated as the least amount of pentosidine significantly different from zero at 95% confidence limits, was 0.05 pmol, and 50% inhibition occurred at 1.8 pmol (Fig. 15). The inventors used the linear portion of the curve (0.1-100 pmol) as the working range. Cross-reactivitv studies. The antibody to pentosidine did not recognize L-lysine, L-arginine, pyridine, or

SUBSTITUTESHEET(RULE25)

guanine, and <0.03% relative cross-reactivity was observed for compounds containing a pyridinium ring and (or) an imidazole ring, e.g. , _/ 3-carboline, quinoline, azabenzim- idazole, or benzimidazole (Table V) . The absence of significant cross-reactivity with pentosidine-like compounds suggested that the antibody is specific to pentosidine.

SUBSTITUTE5HEET(RULE25)

TABLE V

Cross-reactivity of various compounds with anti-pentosidine antibody

* ND, not detected; i.e., > 1 μmol was required for 50% inhibition. Each compound was tested separately at 10 fmol to 1 μmol/well under the standard assay conditions described in the text.

Detection of pentosidine in proteins incubated with ribose. BSA incubated with ribose showed cross-reactivity with inhibition curves parallel to that of standard pentosidine, suggesting that the cross-reactivity was specific for a pentosidine epitope. The cross-reactivity increased as a function of time (Fig. 16) . SDS-PAGE of lysozyme-ribose clearly showed increasing amounts of oligomers formed as a function of time (Fig. 17A) . The

SUBSTITUTESHEET(RULE25)

oligomers were stained by the enzyme-labeled antibody (Fig. 17B) , and the blotting was suppressed when the antibody was preincubated with free pentosidine (Fig.l7C). The monomer also stained slightly even at day 0, but this blotting was not blocked by free pentosidine. Additional studies revealed that this staining was due to the secondary antibody and intrinsic to lysozyme, because it was also observed with enzyme-linked antibody from various vendors. Sample treatment of BSA-ribose and plasma. Incubating plasma protein with 20 g/kg Pronase E at 37°C enhanced the immunoreactivity as a function of time, reaching a plateau by 16 h (Fig. 18) . After 24 h of incubation with enzyme, the digestion (estimated by ninhydrin assay) was 70-84% of that obtained by acid hydrolysis (Moore, S., Stein, W.H., A modified ninhydrin reagent for the photometric determination of amino acid and related compounds. J. Biol. Chem.. 211, pp. 907-13 (1954)). Pronase E digestion of proteins incubated with ribose increased cross- reactivity in the ELISA by -40-600-fold, and the values were close to those obtained by acid hydrolysis (Table VI) . The inhibition curves of digested serum proteins, BSA-ribose, BSA-lysozyme, and pentosidine were parallel to each other (Fig. 19) . Therefore, the inventors used enzyme digestion to analyze pentosidine in plasma proteins.

SUBSTITUTESHEET(RULE25)

TABLE VI

Quantification by ELISA of pentosidine in BSA incubated with 0.1 mol/L D-ribose.

Pentosidine in plasma. To assay pentosidine in plasma, the inventors used an equivalent of 3.125 μL of digested plasmas per well. Because plasma protein concentration varies little (62-77 g/L) , the inventors expressed the pentosidine concentration per sample volume instead of protein content. The ELISA values (y) correlated significantly with values obtained by HPLC (x) : y = 0.94x + 56 (r = 0.98, P <0.01) (Fig. 20). Reproducibility studies at high, middle, and low pentosidine values showed intraassay and interassay CVs of 6.9-10.6% and 10.8-18.79%, respectively (Table VII). Analytical recovery tests showed that 99-101% of added standard pentosidine was recovered (Table VIII) .

SUBSTITUTESHEET(RULE25)

TABLE VII

Reproductibility of pentosidine quantification by ELISA in Pronase E-digested human plasma.

SUBSTITUTESHEET(RULE25)

TABLE VIII

Analytical recovery of standard pentosidine added to digested plasma.

^• (Measured value/calculated value) x 100%. ^b Pentosidine pmol/equivalent of 3.125 μL of digested plasma.

The data in Fig. 21 show that pentosidine is increased in most specimens from diabetic and uremic patients but not in plasma from hospitalized nondiabetic and nonuremic patients. Concentrations (mean + SD) were 1620 + 1940, 2620 + 1310, 145 + 158, and 161 ± 55 μmol/L in diabetic, uremic, nondiabetic/nonuremic, and control subjects, respectively

(P <0.01). The mean concentrations of pentosidine in the diabetic and the uremic subjects were respectively 10- and

17-fold greater than in the controls. Pentosidine in skin collagen. For determination of pentosidine in insoluble skin collagen, the samples were hydrolyzed and an equivalent of 40 μg of collagen was used per well. An age-related increase in skin pentosidine was seen in nondiabetic individuals (Fig. 22) . Again, values from diabetic and uremic subjects were much greater than in age matched normal subjects.

DISCUSSION

The inventors have demonstrated the feasibility of quantifying pentosidine by ELISA, using a polyclonal rabbit antibody against synthetic pentosidine covalently linked to

KLH. The ELISA was as sensitive as the HPLC procedure for quantification of pentosidine in a protein hydrolysate.

However, the ELISA showed somewhat reduced affinity for pentosidine in native protein. Similar problems have been encountered with antibodies to pyrralline, another advanced

SUBSTITUTESHEET(RULE25)

Maillard reaction product (Miyatra, S., Monnier, V.M. Immunocytochemical detection of advanced glycation end product in diabetic tissues using monoclonal antibody to pyrraline. J. Clin. Invest.. 89, 1102-12 (1992)). Two factors may be responsible for the impaired ability of antibodies to haptens to recognize modifications in intact proteins. First, the modification may be inaccessible to the antibody, possibly because of conformational change induced by the modification. The inventors' own data and observations by others have shown that short term preincubation of the modified protein with NaOH is generally accompanied by increased cross-reactivity, possibly related to an unfolding of the proteins (Nakaya a, H. , Mitsuhashi, T. , Kuwajima, S., Aoki, S. , Kuroda, Y., Itoh, T. , et al. Immunochemical detection of advanced glycation end products in lens crystallins from streptozocin-induced diabetic rat. Diabetes. 42, pp.345-50 (1993), Mitsuhashi, T. , Nakayama, H. , Itoh, T., Kuwajima, S., Aoki, S., Atsumi, T. , et al. Immunochemical detection of advanced glycation end products in renal cortex from STZ-induced diabetic rat. Diabetes. 42, pp. 826-32 (1993)). After digestion of BSA-ribose and plasma proteins with Pronase E, the immunoreactivity was increased and became close to that of acid-hydrolyzed samples (Table VI) . Although Pronase E cannot digest serum proteins completely, almost all pentosidine was recovered by the ELISA (Fig.

SUBSTITUTESHEET(RULE25)

19) , which suggests that this pretreatment for plasma protein or BSA-ribose is enough for recognition by the antibody.

The second, and more important, mechanism underlying decreased affinity of antibody for the native hapten is probably steric inhibition by the protein backbone. For both the pyrraline (Miyata, S., Monnier, V.M. Immunocytochemical detection of advanced glycation end product in diabetic tissues using monoclonal antibody to pyrraline. J. Clin. Invest.. 89, pp. 1102-12 (1992); Hayase, F. , Nagaraj, R.H. , Miyata, S., Njoroge, G. , Monnier, V.M. Aging of proteins: immunological detection of a glucose-derived pyrrole formed during Maillard reaction in vivo. J. Biol. Chem.. 264, pp. 3758-64 (1989)) and the pentosidine antibody, the aliphatic chains of e-aminocaproic acid and lysine/arginine, respectively, provided ample spacing to allow the full interaction with the carbodiimide-linked hapten as immunogen. Indeed antibodies to the pyrraline conjugate recognize the side chain of the structure with high specificity and sensitivity. In contrast, several authors have been able to obtain antibodies to advanced glycation end-product proteins after long-term incubation with glucose or glucose 6-phosphate (Nakayama, H. , Taneda, S., Kuwajima, S., Aoki, S., Kuroda, Y. , Misawa, K. , et al. Production and characterization of antibodies to advanced glycation

SUBSTITUTESHEET(RULE25)

products on proteins. Biochem. Biophvs. Res. Commun.. 162, pp. 740-5 (1989); Horiuchi, S., Araki, N. , Morino, Y. Immunochemical approach to characterize advanced glycation end products of the Maillard reaction. J. Biol. Chem.. 266, pp. 7329-32 (1991); Makita, Z., Vlassara, H. , Cerami, A., Bucala, R. Immunochemcial detection of advanced glycosylation end products in vivo. J. Biol. Chem. _r 267, pp. 5133-8 (1992)). These antibodies, directed to still-unknown epitopes, showed increased cross-reactivity with aging and cataractous lens crystallins (Araki, N. , Ueno, N. , Chakabarti, B., Morino, Y., Horiuchi, S. Immunochemical evidence for the presence of advanced glycation end products in human lens proteins and its positive correlation with aging. J. Biol. Chem.. 267, 10211-4 (1992)), lens crystallins from diabetic rats (Nakayama, H. , Mitsuhashi, T. , Kuwajima, S., Aoki, S. , Kuroda, Y., Itoh, T. , et al. Immunochemical detection of advanced glycation end products in lens crystallins from streptozocin-induced diabetic rat. Diabetes. 42, pp.345-50 (1993)), renal tissue from diabetic rats (Mitsuhashi, T., Nakayama, H. , Itoh, T. , Kuwajima, S., Aoki, S., Atsumi, T. , et al. Immunochemical detection of advanced glycation end products in renal cortex from STZ-induced diabetic rat. Diabetes. 42, pp. 826-32 (1993)), aortic collagen from diabetic rats (Makita, Z., Vlassara, H. , Cerami, A., Bucala, R. Immunochemcial detection of advanced

SUBSTITUTESHEET(RULE25)

glycosylation end products in vivo. J. Biol. Chem.. 267, pp. 5133-8 (1992)), serum protein from diabetic humans (Makita, Z., Vlassara, H. , Cerami, A., Bucala, R. Immunochemcial detection of advanced glycosylation end products in vivo. J. Biol. Chem.. 267, pp. 5133-8 (1992)), and hemoglobin from diabetic humans (Makita, Z., Vlassara, H. , Rayfield, E. , Cartwright, K. , Friedman, E. , Rodby, R. , et al. Hemoglobin-AGE: a circulating marker of advanced glycation. Science. 258, pp. 651-3 (1992)). Preliminary studies by Makita et al. (Makita, Z., Vlassara, H. , Rayfield, E. , Cartwright, K. , Friedman, E. , Rodby, R. , et al. Hemoglobin-AGE: a circulating marker of advanced glycation. Science. 258, pp. 651-3 (1992)) have shown that pentosidine does not cross-react with the advanced glycation end-product antibody, suggesting that pentosidine is one of several protein modifications resulting from the Maillard reaction in vivo. Most in vivo studies so far indicate that the pattern of pentosidine formation in collagen is similar to that of carboxymethyllysine, another advanced Maillard reaction and glycoxidation product of Arnadori products (McCance, D.R. , Dyer, D.G., Dunn, J.A. , Bailie, K.E., Thorpe, S.R., Baynos, J.W. Maillard reaction products and their relation to complication in insulin-dependent diabetes mellitus. JX. Clin. Invest.. 91, pp. 2470-8 (1993)).

SUBSTITUTE SHEET (RULE 25)

From a clinical viewpoint, quantification of pentosi¬ dine (or carboxymethyllysine) in skin collagen appears to provide a more reliable measure of cumulative glycemia over several years. In contrast, glycated collagen itself, i.e., the Amadori product is in steady state, and its concentrations in collagen have been slowly reversible in diabetic subjects that were tightly controlled for 4 months (Lyons, T.J., Bailie, K.E., Dyer, D.G., Dunn, J.A. , Baynes, J.W. Decrease in skin collagen glycation with improved glycemic control in patients with insulin-dependent diabetes mellitus. J. Clin. Invest.. 87, pp. 1910-5 (1991) ) . Further studies have documented a relation between skin collagen concentrations of glycoxidation products and severity of retinopathy (McCance, D.R., Dyer, D.G., Dunn, J.A. , Bailie, K.E., Thorpe, S.R. , Baynos, J.W. Maillard reaction products and their relation to complication in insulin-dependent diabetes mellitus. J. Clin. Invest.. 91, pp. 2470-8 (1993)), Beisswenger, P.J., Moore, L.L., Brink- Johnsen, T. Increased collagen-linked pentosidine levels and advanced glycosylation end products in early diabetic nephropathy. J. Clin. Invest.. 92, pp. 212-7 (1993)), nephropathy (McCance, D.R., Dyer, D.G., Dunn, J.A. , Bailie, K.E., Thorpe, S.R., Baynos, J.W. Maillard reaction products and their relation to complication in insulin-dependent diabetes mellitus. J. Clin. Invest.. 91, pp. 2470-8 (1993)), Beisswenger, P.J., Moore, L.L., Brink-Johnsen, T.

SUBSTITUTE SHEET (RULE 25)

Increased collagen-linked pentosidine levels and advanced glycosylation end products in early diabetic nephropathy. J. Clin. Invest.. 92, pp. 212-7 (1993)), and overall severity of complications (Sell, D.R., Lapolla, A., Odetti, P., Fogarty, J. , Monnier, V.M. Pentosidine formation in skin correlates with severity of complication in individuals with long-standing IDDM. Diabetes. 41, pp. 1286-92 (1992)). Thus, quantification of tissue pentosidine might help evaluate the risk of developing severe com- plications in young subjects with high skin concentrations of pentosidine.

Finally, an antibody to pentosidine is helpful for in situ localization of advanced Maillard reaction products in tissues in which long-lived proteins undergo aging. For example, a study with the antibody the inventors raised in rabbits revealed strongly pentosidine immunoreactivity in brain neurofibrillary tangles from the brains of patients with Alzheimer disease (Smith, M.A. , Taneda, S., Richey,

P.L. , Miyata, S., Yan, S.D., Stern, D., et al. Advanced Maillard reaction endproducts are associated with Alzheimer disease pathology. Proc. Natl. Acad. Sci. USA. 91, pp.

5170-4 (1994)) . Although it sense unlikely that Alzheimer disease would be related to a systemic Maillard reaction, immunohistochemical studies with antibodies to pentosidine and other Maillard reaction products are needed to assess damage to cellular and extracellular proteins in aging and

SUBSTITUTESHEET(RULE25)

age-related diseases. Preliminary studies in that direction have revealed pentosidine immunoreactivity in the arterial wall and extracellular matrix of heart, kidney, lung, skin, and dura mater obtained at autopsy. However, because of the lower affinity of our antibody for protein- linked pentosidine, immunocytochemistry with this antibody will significantly underestimate true concentrations of tissue pentosidine.

In summary, the development of a polyclonal antibody to pentosidine will allow investigators to make rapid serial quantifications of pentosidine in a large number of hydrolyzed proteins. This new tool is expected to be helpful for assessment of the advanced Maillard reaction in vivo.

Example 3

This example relates to a process for accessing aging in mammals. The process involves quantitating the amount of pentosidine present in mammalian tissue.

MATERIALS AND METHODS Tissue Procurement. Skin samples (Fig. 23) were obtained from human at autopsy (Institute of Pathology, Cleveland, OH) ; Hormel breed of miniswine (Dr. Wesley Johnson, Laurel, MD) and the least shrew (from Dr. Orin Mock, Kirksville, MO) at necropsy; rhesus monkey (from Dr.

SUBSTITUTESHEET(RULE25)

Karen Reiser, Davis, CA) , cow (Dr. Herdt, Michigan State University, East Lansing, MI) and 2 year-old beagle dogs (Dr. Sanford Emery, Dept. of Surgery, Case Western Reserve University) by biopsy. All other beagle dog skin samples were obtained by Drs. Mark Lane and George Roth by biopsy at the Gerontology Research Center, Baltimore, MD.

Monkey skin samples of Fig. 25 were obtained by Drs. Mark Lane and George Roth by biopsy from rhesus and squirrel monkeys house at the NIH Animal Center at Pooleville, MD. First samples were collected between October and November 1990, while the second samples were collected between December 1992 and January 1993, In most instances, samples were collected from the same monkey at both time periods (squirrel, n=5; rhesus, n=8) . However, due to animal death or other circumstance, some samples were collected only once either during the first (squirrel, n=l; rhesus, n=3) or the second (squirrel, n=4; rhesus, n=5) period. A total of 10 squirrel and 16 rhesus monkeys are represented. Procedures for husbandry of these monkeys have been previously published by Ingram et al. (Ingram, D.K., Cutler, R.G., Weindruch, R. , Renquist, D.M. , Knapka, J.J. , April, M. , Belcher, C.T. , Clark, M.A. , Hatcherson, CD., Marriott, B.M., Roth, G.S. J. Gerontol. 45, pp. B148- B163 (1990). Tail tendons and dorsal skin of Fig. 26 were obtained from ad libitum and dietary restricted Fischer 344 rats by

SUBSTITUTE SHEET(RULE25)

Dr. Edward Masoro, San Antonio, TX. Husbandry of these rats has been previously described by Masoro et al. (Masoro, E.J. , McCarter, R.J.M. , Katz, M.S. & McMahan, CA. J. Gerontol. 47, pp. B202-B208 (1992)). Processing of samples. Skin samples were minced and homogenized in phosphate buffer saline (PBS), pH 7.4. The insoluble collagen was recovered by centrifugation (DuPont Sorvall, Wilmington, DE) at 10,000 X g for 20 minutes and subsequently sequentially extracted for 24 hrs at 4°C in each of 2:1 chloroform:methanol, 1 M sodium chloride (pH 7.4), 0.5 M acetic acid and 0.1 mg/ml pepsin in 0.5 M acetic acid. The insoluble collagen was washed three times in water and freeze dried.

Tendons were carefully dissected from skinned tails and washed in PBS. Tail tendon breaking time was subsequently measured as previously described by the inventors in Monnier et al. (Monnier, V.M. , Sell, D.R. , Abdul-Karim, F.W., & Emancipator, S.N. Diabetes. 37, pp. 867-872 (1988)). Weights of the equivalent of 270 to 280 mg were attached to tendons with 4-0 silk suture (Ethicon) . Tendon and weight were immersed in 7 M urea kept at 40+0.5°C, and the other end was attached to a hook connected to an electric time. When the tendon broke, the hook was released, and the timer stopped.

SUBSTITUTESHEET(RULE25)

A further piece of tendon was excised from the tail for pentosidine analysis. The tendon was washed in PBS, extracted in 2:1 chloroform:methanol, and freeze dried.

Preparation of Tissue Hydrolysate. A total of 5 mg skin was placed into a 13 X 100 mm screw-capped tube. A total of 2 ml deaerated 6 M HCl was added to the tube. The tube was purged with nitrogen and sealed with a Teflon- faced rubber-lined cap. All samples were acid-hydrolyzed for 24 hr at 110°C The acid was evaporated by a Speed-Vac (Savant Instruments, Inc., Farmingdale, N.Y.) and each sample was reconstituted with 1 ml of water containing 0.01 M heptafluorobutyric acid (HFBA) . After filtering each sample with a 4 μm filter, collagen content was determined for all samples by the hydroxyproline colorimetric assay assuming a collagen content of 14% cydroxyproline by weight as previously described (Sell, D.R. & Monnier, V.M. J. Biol. Chem.. 264, pp. 21597-21602 (1989). In any given experiment, all samples were equalized for hydroxyproline (approximately 250 μg/ml) by diluting the sample with water containing 0.01 M HFBA.

High Performance Liouid Chromatography fHPLC . Pentosidine was determined by a repetitive injection technique as previously described by Odetti et al. (Odetti, P., Fogarty, J. , Sell, D.R. & Monnier, V.M. Diabetes. 41, pp. 153-159 (1992)). Samples of approximately 100-200 μl volumes equivalent to 45 μg of hydroxyproline were injected

SUBSTITUTESHEET(RULE25)

onto a C-18 reverse-phase analytical column (Vydac 0.46 X 25 cm, 10 μm; The Separations Group, Hesperia, CA) . The column was attached to a High Performance Liquid Chromatograph (HPLC) equipped with model 510 pumps, a model 712 WISP automatic injector, and a model 680 controller (Waters, Milford, MA) . The pentosidine peak was monitored at Ex 335/Em 385 nm by an on-line JASCO 821-FP spectrofluorometer (JASCO Incorporated, Easton, MD) . The eluent from 29 to 34 min was collected, dryed and reconstituted in 200 μl of 0.02 M sodium acetate (ph 4.47) and in turn 160 μl of this volume were injected into a sulfopropyl cation-exchange column (Protein-Pak, SP 5PW, 7.5 X 75 mm, Waters) . The column was eluted with a gradient of 0 to 0.06 M NaCl in 0.02 M sodium acetate over 40 minutes at a flow rate of 1.0 ml/min. The fluorescence detector was interfaced to a computer loaded with Borwin Chromatography Software for recording and integration of chromatographic peaks (JMBS Developpements, Le Fontanil, France) . Pentosidine eluted approximately at 18 minutes. Statistical Analyses. For each response function of Fig. 23, a regression lines (SPSS Inc., Chicago, IL) was computed using the square root transformation of data points (Neter, J. & Wasserman, W. Applied Linear Statistical Models, pp. 131-136, 273-296 (Richard D. Irwin, Inc., Homewood, IL) (1974)). A confidence interval was similarly computed using the error of prediction formula

SUBSTITUTESHEET(RULE25)

(Steel, R.G.D. & Torrie, J.H. Principles and Procedures of

Statistics, pp. 471, 542-543 (McGraw-Hill Inc., New York)

(1980)). The results of these analyses (Figs. 23, Table X) are expressed in their untransformed forms as second-order polynomial response functions (i.e., quadratic equations)

(Neter, J. & Wasserman, W. Applied Linear Statistical

Models, pp. 131-136, 273-296 (Richard D. Irwin, Inc.,

Homewood, IL) (1974) ; Steel, R.G.D. & Torrie, J.H.

Principles and Procedures of Statistics, pp. 471, 542-543 (McGraw-Hill Inc., New York) (1980)). In cow skin (Fig.

23) , one data point was found to reside from all other points, however, it could not be statistically proven by a test (Snedecor, G.W. & Cochran, W.G. Statistical Methods.

6th Ed., pp. 157-158 (The Iowa State University Press, Ames, Iowa) (1967)) to be an outlier, thus the regression line was determined for all data points.

SUBSTITUTESHEET(RULE25)

TABLE IX

Reported longevities and glycemic levels of animals used in experiments.

* Except for glycemic values in the dog and cow, all values have been determined in the animal cohor from which the tissue was obtained.

⁺ Dietary restricted rats are given in parenthesis.

SUBSTITUTE SHEET (RULE 25)

TABLE X

Regression equations for data presented in Figs. 23-24. 26 and 27.

en * x: age in years, y: pentosidine in pmol/mg collagen.

* Squirrel monkey data of Fig. 25 were remodeled from a linear (R ² =0.71l) to a quadratic (R ² =0. function.

In the analysis of Fig. 25, biopsies were taken from squirrel and rhesus monkey at two different sampling periods, 1991 and 1993 (biopsy time) . In half of the total

26 monkeys represented, the biopsy was taken from the same monkey at both time periods, while in the other monkeys, a single biopsy was collected either in 1991 or 1993

(replication) . Thus, the regression analysis consisted of pentosidine levels versus age, monkey, biopsy time, and replication. The regression equations are expressed in pentosidine versus age for each monkey species. The other factors (biopsy time, replication) were pooled into the constant of the equation. The comparison of regression lines for slope differences was made according to the test given by Snedecor and Cochran (Snedecor, G.W. & Cochran, W.G. Statistical Methods. 6th Ed., pp. 157-158 (The Iowa

State University Press, Ames, Iowa) (1967)).

Means of Fig. 26 were compared by the Student-Newman-

Keuls multicomparison test (Steel, R.G.D. & Torrie, J.H.

Principles and Procedures of Statistics, pp. 471, 542-543 (McGraw-Hill Inc. , New York) (1980)) following the analysis of variance (SPSS Inc., Chicago, IL) .

In order to make a comparison of pentosidine rate of formation in skin among different animals of Figs. 23, 25 and 26, regression equations were modeled as quadratic (parabolic) response functions (Table X) . This involved remodeling squirrel monkey data of Fig. 25 from a linear to

SUBSTITUTESHEET(RULE25)

a quadratic function which slightly changed the coefficient of determination from R ² =0.71l to 0.706, respectively (Table X) . Furthermore, skin levels of ad libitum and dietary restricted rats of Fig. 26 were modeled with quadratic equations (Table X) .

RESULTS Pentosidine was determined in skin collagen from different species of animals varying in ages (Fig. 23) and longevities (Table X) . Graphic presentations of these results are shown separately for each animal species in Fig. 23 and presented compositely for all animals in Fig. 24.

Highly significant (P<.0001) increases with age were found in all species presented in Fig. 23. Best-fit regression analysis showed curvilinear models for the shrew

(R ² =0.915), dog (R ² =0.872), and cow (R^O.851), linear models for the pig (R ² =0.975) and monkey (R ² =0.569) and an exponential model for the human (R=0.863). However, for the later statistical comparisons, all models presented are curvilinear (Figs. 23 & 24) represented by quadratic equations (Table X) . These changes in models for the pig, monkey and human had slight effects on their coefficients of determination (R=0.966, 0.568, 0.860, respectively) meaning they were equally appropriate in explaining total variation of data.

SUBSTITUTE SHEET(RULE25)

In the analyses of pentosidine level as a function of age for given animal species, absolute level and the rate of formation were the two most important factors in explaining levels in skin. However, these two factors were not equivalent in explaining pentosidine relationship to life span. The most rapid increase was noted for the least shrew (Fig. 23), a very small animal, weighing only 4-5 g as an adult, belonging to the order Insectivora and the family Soricidae (Mock, O.B. Lab. Anim. Sci.. 32, pp. 177- 179 (1982), Mock, O.B. AALAS Bulletin, p. 35 (July 30, 1991), Mock, O.B. The Gerontoloσist. 33, pp. 169 (October Special Issue 1) (1993)). In this species, absolute levels present in skin was small in comparison to all other animal species, reaching approximately 7 pmol/mg collagen at 3 years (Fig. 23) . Levels in miniature swine skin (Fig. 23) increased rapidly as well, reaching approximately 15 pmol/mg collagen at 14 years. This increase is much more rapid than that noted for rhesus monkey (Fig. 23) where levels approximated 20 pmol/mg collagen at 26 years and whose life span is greater than shrew or miniature swine (Table IX) . Highest absolute levels were noted for cow, dog, and human skin collagens (Fig. 23) . In cow skin, levels increased very rapidly reaching approximately 30 pmol/mg collagen at 9 years (Fig. 23) . Surprisingly, levels for dog skin approach 60 pmol/mg collagen at 14 years which is much higher than anticipated. The highest

SUBSTITUTESHEET(RULE25)

absolute levels in the study were observed for human skin, reaching approximately 100 pmol/mg collagen at 92 years (Fig. 23).

The shrew study is of particular interest because these animals have a high resting metabolic rate (Mock, O.B. Lab. Ani . Sci.. 32, pp. 177-179 (1982), Mock, O.B. AALAS Bulletin, p. 35 (July 30, 1991), Mock, O.B. The Gerontoloσist. 33, pp. 169 (October Special Issue l) (1993)). If the regression line observed for this species is extrapolated beyond 3.5 years, the projected rate of increase in levels occurs faster than levels observed for the dog or cow of comparable ages, and much faster than rates observed for the pig, monkey or human (Fig. 24) . Further evidence for the validity of this observation comes by noting differences in pentosidine formation rates as reflected by the coefficients of x2 (curvature effect coefficients) for the determined regression lines of Fig. 23 & 24. As shown in Table X, the magnitude of these coefficients is greatest for the shrew. Levels were also studied in extracted skin collagen samples from two different species of monkey with different longevities (Fig. 25, Table IX) . In independent analysis by species, age was significant for both squirrel (P<.0004) and rhesus (P<.015) monkey. In combined regression analysis of both monkey species, age was again the most important factor in explaining increased levels (P=0.0002) .

SUBSTITUTESHEET(RULE25)

Slope analysis showed that the rate of increase in levels for the shorter-lived squirrel monkey was significantly (P<.001) greater than that noted for rhesus monkey (Fig.

25) . Pentosidine levels in skin and tail tendons of Fischer 344 rats were studied in relationship to tail tendon breaking time, a parameter of collagen crosslinking (Fig.

26) . The effects of dietary restriction on tendon breaking time and pentosidine were also studied. The results show that tendon breaking time significantly (P<.0001) increased with age in both ad libitum and dietary restricted rats (Fig. 26A) . The rate of increase occurred much more rapidly in an libitum-fed compared with dietary restricted rats (P<.0001) . Measured tendon breaking times at all ages measured were significantly (P<.01) greater for ad libitum compared with dietary restricted rats (Fig. 26A) .

Likewise, the same trends noted for tendon breaking time also occurred for tendon (Fig. 26B) and skin pentosidine levels (Fig. 26C) except more variation was observed. Levels in both tissues increased significantly (P<.0001) with age in both ad libitum and dietary restricted rats (Fig. 26B, 26C) and again the rate of increase occurred faster for ad libitum compared with dietary restricted animals (tendon, P<.035; skin, P<.001). Differences in levels between ad libitum and dietary

SUBSTITUTESHEET(RULE25)

restricted rats were significant (P<.05) at ages greater than 12 months.

In Fig. 27, the coefficient of x ² for each equation of Table X, representing the quadratic rate component of pentosidine formation in skin, was plotted versus the maximum life span of the respective animal species. The results showed a significant (P<.004) inverse relationship between the rate of formation and maximum life span (Fig. 27A) . Rates of formation were highly accelerated in short- lived species like the rat and shrew in comparison to longer-lived species such as the monkey or human. Differences between shorter and longer-lived species were such that the rate of formation was significantly (P<.004) and inversely related to the logarithm of maximum life span as observed by the linear line drawn through the points of Fig. 27B. Dietary restriction, notably known for its life span extendibility, markedly decelerated the rate of formation in the Fischer 344 rat (Fig. 27A, 27B) .

DISCUSSION The results of this example show that glycoxidation rate as measured by pentosidine increases as a function of age in skin collagen of different mammalian species. Obviously, this relationship is not absolute in the sense that maximum levels observed varied within species of approximately the same life span (dog, pig, cow) . However,

SUBSTITUTESHEET(RULE25)

the finding of an inverse relationship between glycoxidation rate and longevity within the same animal family, squirrel monkey versus rhesus monkey (Fig. 25) , and among different animal species (Fig. 27) , supports the notion that a process which is reflected in pentosidine formation may play a role in determination of longevity. Thus, pentosidine accumulation rates may be controlled by factors inherited to species specific skin biology such as collagen turnover and/or the ability to withstand the potentially harmful effects of glycoxidation due to the Maillard reaction.

Although it is conceivable that collagen turnover may be regulated independently from glycoxidation, glycoxidation and crosslinking themselves could impair collagen turnover since collagen from both old and diabetic individuals are pentosidine-rich and less susceptible to collagenase digestion (Kohn, R.R. Principles of Mammalian

Aging, pp. 1-229 (Prentice-Hall, Englewood Cliffs, N.J.)

(1978) ; Kohn, R.R. Testing the Theories of Aσinσ. pp. 221- 231, eds. Adelman, R.C. & Roth, G.S. (CRC Press, Inc., Boca

Raton, FL) (1982) ; Monnier, V.M. The Maillard Reaction in

Aσinσ. Diabetes, and Nutrition, pp. 1-22, eds. Baynes, J.W.

& Monnier, V.M. (Alan R. Lliss, New York) (1989); Sell,

D.R. & Monnier, V.M. J. Biol. Chem.. 264, pp. 21597-21602 (1989)). Thus, a vicious cycle consisting of glycation- mediated crosslinks may impair turnover itself. Another

SUBSTITUTESHEET(RULE25)

factor which may be involved in controlling glycoxidation rates is glycemia. Cerami (Cerami, A. J. Am. Geriatr.

Soc.. 33, pp. 626-634 (1985)) hypothesized that glucose may serve as a mediator of aging based on findings that complications of aging such as increased stiffening of joints, arteries, long and heart, bone loss, loss of lens accommodation, cataract formation, atherosclerosis and cardiovascular disease, are dramatically worsened by diabetes. Indeed, these sequelae of aging as well as glycoxidation rates occurs twice faster in diabetes

(Monnier, V.M. The Maillard Reaction in Aσinσ. Diabetes. and Nutrition, pp. 1-22, eds. Baynes, J.W. & Monnier, V.M.

(Alan R. Lliss, New York) (1989); Monnier, V.M. , Sell,

D.R. , Miyata, S. & Nagaraj, R.H. The Maillard Reaction in Food Processing. Human Nutrition, and Physiology, pp. 393-

413, eds. Finot PA, Aeschbacher HU, Hurrell RF & Liardon R

(Birkhauser Verglag, Basel, Switzerland) (1990) ; Monnier,

V.M. , Sell, D.R., Nagaraj, R.H. & Miyata, S. Gerontology.

37, pp. 152-165 (1991)), a syndrome of accelerated aging (Hamlin, C.R. , Luschin, J.H. & Kohn, R.R. Diabetes, 24, pp.

902-904 (1975)).

In aging studies, however, glycemia and glycation are not strictly related. First, there is no or only a weak correlation between glycemia and species longevity (Table IX). In most mammalian species, for example, glycemia is similar and does not correlate inversely with longevity.

SUBSTITUTESHEET(RULE25)

Some birds, for example, have very high glycemia (Homles, D.J. & Austad, S.N. J. Gerontol. 50A, pp. B59-B66 (1995)) and yet live for decades. Such high glycemic levels suggest that defense mechanisms must exist to circumvent the potentially harmful effects of glycoxidation (Hamlin, C.R., Luschin, J.H. & Kohn, R.R. Diabetes. 24, pp. 902-904 (1975) . Secondly, glycation of long-lived proteins such as collagen increases only slightly with age (Dunn, J.A. , McCance, D.R. , Thorpe, S.R., Lyons, T.J. & Baynes, J.W. Biochemistry. 30, pp. 1205-1210 (1991)). The possibility remains, however, that some of the differences in the accumulation of glycated long-lived proteins are species or tissue-specific. Thirdly, glucose intolerance generally increases with age and may be associated with glycohemoglobin levels (Shimokata, H. , Muller, D.C., Fleg, J.L. , Sorkin, J. , Ziemba, A.W. & Andres, R. Diabetes. 40, pp. 44-51 (1991)), but age-related processes such as coronary heart disease still proceed in individuals who are not glucose intolerant (Schildkraut, J.M. , Myers, R.H. , Cupples, A., Kiely, D.K. & Kannel, W.B. Am J. Cardiol. 64, pp. 555-559 (1989)).

Albeit the lack of a strict relationship between aging and glycemia, the latter is expected to play some role for longevity, perhaps in the form of modulation of the aging rate. Data by Masoro et al. (Masoro, E.J. , McCarter,

R.J.M. , Katz, M.S. & McMahan, CA. J. Gerontol. 47, pp.

SUBSTITUTE SHEET (RULE 25)

B202-B208 (1992)), studying the mode of action of dietary restriction in extending life span of the Fischer 344 rat, showed that these rodents had lower mean 24-hour plasma glucose and lower plasma insulin levels throughout their life span compared with the ad libitu -fed controls. Surprisingly, oxygen consumption, respiratory quotient and glucose utilization rates were similar in ad libitum and dietary restricted groups. The present results showed that dietary restriction significantly reduced both tail tendon breaking time, a parameter of crosslinking, and pentosidine in tendon and skin (Fig. 26) .

The mechanism by which dietary restriction acts is unclear, but may involve free radical-mediated damage since both glycoxidation, measured by pentosidine (Fig. 26) , and lipid peroxidation (Yu BP, Lee D-W & Chio J-H. Biological

Effects of Dietary Restriction, pp. 191-197, ed. Fishbein

L. (Springer-Verlag, New York), (1991)) are decreased.

Much interest in recent years has focused on the potential of glycated proteins as a source of free radicals (Kristal, BS & YU, BP. J. Gerontol. 47, pp. B107-B114 (1992)) whereby such free radicals could stem from oxidative reactions of glucose as proposed by Wolff (Wolff, S.P., Jiang, Z.Y. &

Hunt, J.V. Free Rad. Biol. Med.. 10, pp. 339-352 (1991)), the Schiff base adduct (Hayashi, T. & Namfki, M. , Dev. Food Sci.. 13, pp. 29-38 (1986)) and the Amadori product

(Baynes, J.W. , Diabetes. 40, pp. 405-412 (1991); Monnier,

SUBSTITUTESHEET(RULE25)

V.M. , Sell, D.R., Miyata, S. & Nagaraj, R.H. The Maillard Reaction in Food Processing. Human Nutrition, and Physiology, pp. 393-413, eds. Finot PA, Aeschbacher HU, Hurrell RF & Liardon R (Birkhauser Verglag, Basel, Switzerland) (1990). Studies by Fu et al. (Fu, M-X, Thurpe, S.R. & Baynes, J.H. Maillard Reactions in Chemistry. Food, and Health, pp.95-100, eds. Labuza TP, Reineccius GA, Monnier VM, O'Brien JR, Baynes JW. (The Royal Society of Chemistry, Cambridge, U.K.) (1994)) showed that the addition of metal chelators, sulfhydryl compounds and antioxidants to rat tail collagen incubated with glucose inhibited glycoxidation and crosslinking of collagen, but had negligible effects on glycation. Such findings suggest that the permanent damaging effects of glycoxidation are not due to glycation per se, but to subsequent free radical mediated oxidative modifications of long-lived proteins.

Although previous intervention studies with the use of antioxidants were ineffective at prolonging maximum life span (Schneider, E.L. & Reed, J.D. Handbook of the Biology of Aging, pp. 45-76, eds. Finch CE & Schneider EL (Van Nostrand Reinhold Co., New York) (1985)), recent studies with Drosophila transgenic for superoxide dismutase and catalase (Orr, W.C. & Sohal, R.S. Science. 263, pp. 1128-1130 (1994)) had a 14 to 34% increase in maximum life span, implying therefore oxygen free radicals in longevity.

SUBSTITUTESHEET(RULE25)

It is believed that future studies on the prevention of the harmful effects of glycoxidation on long-lived proteins will be achieved through pharmaceutical and genetic interventions. Recently, an attempt has been made to use aminoguanidine to inhibit glycoxidation and crosslinking in animal studies (Vlassara, H. The Gerontologist. 34, pp. 235-236 (October Special Issue 1) (1994)). In diabetes, aminoguanidine has been shown to inhibit basement membrane modifications, atherosclerosis, and kidney disease. In aging studies, although longevity data are not reported, aminoguanidine prevented vasodilatory impairment and albuminuria associated with renal insufficiency in old rat (Vlassara, H. The Gerontologist. 34, pp. 235-236 (October Special Issue 1) (1994)). Thus, whether or not aminoguanidine will be able to extend maximum life span, it is expected to be an useful drug for the treatment of age- related complications. Likewise, another potentially way to intervene against the Maillard reaction would be to genetically engineer animals or cells with enzymes reversing nonenzymatic glycation in vivo. Such enzymes have been recently found in soil organisms (Gerhardinger, C. , Marion, M.S., Rovner, A., Glomb, M. & Monnier, V.M. J. Biol. Chem.. 270, pp. 218-224 (1995).

In short, this example is, to inventor knowledge, the first to document an inverse relationship between species longevity and glycoxidation rates. The mechanism

SUBSTITUTESHEET(RULE25)

underlying this association and its relevance in determining longevity remains to be established.

The invention has been described with reference to the preferred embodiment. Obviously, modifications and alterations will occur to others upon reading and understanding the preceding detailed description. It is inteded that the invention be construed as including all such alterations and modifications insofar as they come within the scope of the appended claims or the equivalents thereof.

SUBSTITUTESHEET(RULE25)