Role of the Maillard reaction in aging and age-related diseases. Studies at the cellular-molecular level

Introduction

It has to be stated from the beginning that the regulation of longevity and the mechanisms of aging are separate processes and have to be studied with this understanding. Longevity determinants are, not entirely but to a significant degree, genetically determined ([1, 2] for review). Aging, however, is mostly post-genetic, partially epigenetic [3] and to a large extent dominated by post-genetic mechanisms [4]. Table 1 gives a non-exhaustive list of the most studied aging mechanisms which were explored at the cellular-molecular level. Although aging itself and age-related diseases are frequently confounded, the underlying mechanisms can be separated in most instances. Maillard reaction, listed on Table 1, is an important contributor both to age-dependent decline of tissue structure and function as well as to several age-related diseases. This is the motivation for its intensive study over the last decades. This trend will certainly continue for the next decades as shown by the number and quality of presentations at the last meeting in Nancy for the centenary celebration of the publication by Maillard concerning his reaction, where an important generation of young scientists presented most of the contributions.

Aging in spare parts

In sharp contradiction to popular belief, vital functions do not decline at the same rate with age. An important contribution to this subject was the listing by Dr Weale in London (UK) of the speed of decline of a number of functions with age [11]. These declines, listed in a table by Weale [11] can be plotted on a graph and extrapolated to the intersection of the graphs with the abscissa representing the age where zero function is left. It can be seen from such a representation (Table 2) that complete loss of function occurs fastest for several elastic tissue properties. The best known being the lens capsule, cataract operations became frequent from 60 years on. The slowest decline in Weale’s table is for the speed of nervous conduction, its decline reaching the abscissa at about 400 years [11]. All other functions decline with speeds in-between these extremes. A caveat to this schematization is the large individual variation for the decline of any of the recorded functions. This is valid also for the age of onset for age-related pathologies where environmental factors and behavior patterns (alcohol, tobacco, obesity) as well as genetic dispositions play a crucial role, as well as the interaction of environmental factors with gene functions [12].

Experimental approach

Over the years we carried out studies on aging and age-related pathologies, mostly at the cellular-molecular level. Cell functions explored comprised entrance in the senescent state, cell death, free radical (or better ROS) production, regulation of the expression of proteolytic activity involved in tissue degradation, and the study of gene expression profiles using microarrays. The role of the Maillard reaction in the above listed processes was explored by adding glycation-products (GPs) to cells and monitoring induced modifications in cell function or gene expressions. The preparation of the GPs, the source of cells and details of culture conditions were described in detail [13].

Age-dependent cell loss by entering senescence and cell death

These studies were based on the paradigm of cell aging as established by Hayflick [14]. Entrance of cells in the senescent state was quantified by the specific staining of senescent cells using the procedure described by the team of Judith Campisi [15], counting cells stained for the expression of SA-β-galactosidase (senescence associated β-galactosidase). Cell death was quantified by counting dead cells floating on top of the culture medium in multiwell plates in standardized conditions [13].

As shown on Table 3, the addition of GPs to cultures of human skin fibroblasts strongly increased cell death. The degree of increase varied according to the GPs used. All the GPs tested increased cell death to variable extents. All these effects were significant and reproducible.

An important observation, relevant for the mechanism of action of GPs was the demonstration that the tendency to increased cell death transmitted to future generations of serially cultured cells, with washed cells in fresh medium without added GPs. Here again, the lysine-glucose generated GPs were the most efficient in transferring increased decay to future generations. The rate of cell proliferation was also increased by most GPs tested [13]. Cell death was of a necrotic type, by lesion of the cell membrane. This observation suggested a possible role for ROS-generation by GPs, demonstrated by several authors [16, 17] and studied in our laboratory also as will be described later.

Table 4 shows the acceleration by GPs of the entrance of cells in the senescent state. We could show that SA-β-gal staining increased linearly with time spent in culture by the cells as well as with passage number by using the Hayflick method for in vitro aging [18].

Matrix degradation by free radical generation

In order to test directly the role of GPs in the degradation of macromolecules of the extracellular matrix (ECM) we used hyaluronan as the target. This polysaccharide is present in most connective tissues and plays an important role by stabilizing and crosslinking proteoglycans in cartilage and other tissues [19]. We could show that hyaluronan is exquisitely sensitive to free radical produced degradation by recording with time its viscosity in a coaxial rotating viscometer [17]. As shown on Figure 1, the addition of a well-known free radical generating reagent, described by Udenfriend et al. [22], comprising 48 mM ascorbic acid, 28 mM EDTA and 5 μM FeCl₂, produced a rapid decrease of viscosity, even at a dilution of 1–1000. GPs added to hyaluronan also produces a decrease of viscosity, comparable to the decrease produced by 1 U/mL testicular hyaluronidase preparation (Figure 2). We could further confirm the GPs-mediated degradation of this polysaccharide by determining the number average molecular weight (Mn) of the hyaluronan preparation exposed to the GPs. Mn for GPs-treated hyaluronan as determined by size exclusion chromatography (SEC) was 137 kDa as compared to the value of native hyaluronan of 582.5 kDa [17]. This degradation could be efficiently inhibited by free radical scavenging reagents as shown on Figure 3. Uric acid proved to be the most efficient inhibitor, followed by deferoxamine and SOD [20]. These experiments also suggested an important role of GPs in osteoarticular diseases because of the crucial role of hyaluronan in cartilage stability [19]. Hyaluronan is, however, present in all tissues and local concentrations were shown to decrease with age [21]. This decrease is presumably a combined result of decreased biosynthesis and increasing degradation.

Figure 1

Demonstration of free radical production by GPs as indicated by the degradation of hyaluronan recorded by a rotating coaxial viscosimeter [17, 20].

Abscissa: time, minutes; Ordinates: viscosity in centipoises; Modified from Deguine et al., [21]. ♦ Hyaluronan in PBS; ● Glycated lysine at two concentrations, 5.55 mM, upper curve, 16.65 mM, lower curve; ▴ Udenfriend reagent, diluted 1/1000.

Figure 2

Inhibition of free radical-mediated degradation of hyaluronan by scavengers.

Abscissa: % inhibition. Top-3 columns: GP added: lysine–glucose 16.65 mM. Bottom-3 columns: GP added: the same as above 5.5 mM. ░ SOD 600 U/mL; ▒ Deferoxamine 10 μM; 3×10⁴ cells per mL; ▄ Uric acid 5 mM. Modified from: Deguine [17] and Deguine-Delay [20].

Figure 3

Upregulation of elastase type endopeptidase activity by GPs added to human skin fibroblast cultures, as determined with a synthetic specific substrate, N-succinyl-(ala)₃pNa. GPs added were: glycated arginine; glycated BSA in presence of FeCl₂; glycated lysozyme; glycated BSA with methylglyoxal. For details the original publications can be consulted [23]. Ordinates: increase of elastase activity as compared to control taken as 100% (modified from ref. [23]).

Role of the Maillard reaction in atherogenesis and heart failure

Cardiovascular diseases are certainly multifactorial [24]. GPs were shown to intervene by several mechanisms claimed to be involved in the generation of these pathologies. Our attention was directed to the role of ECM-components and especially of their age-dependent modifications in these pathologies. Since the isolation of the first elastin-degrading enzyme by Balo and Banga, these authors proposed an important role for elastin degradation in the initiation of the atherosclerotic lesions [25]. We could show that vascular smooth muscle cells (SMCs) produce locally elastolytic enzymes shown to be essentially MMP-2 and MMP-9 [6, 7]. Several other enzymes brought to the lesion by migrating monocytes also play a role. We could show by ELISA-titration on more than 2000 blood samples that elastin peptides derived from the degradation of elastic fibers are present in the blood circulation [26]. These peptides act as agonists of the elastin-receptor present on the membrane of SMCs [27]. Among other reactions they trigger the release of free radicals and elastolytic endopeptidases [28]. As shown in Figure 3, GPs added to fibroblast cell cultures induce an increased production of elastase activity. Therefore, GPs contribute to elastin degradation amplified by the elastin receptor. Elastin peptides added to lymphocyte cultures were shown to produce apoptotic cell death [29].

Another aspect of the role of Maillard reaction in atherogenesis is attributed to collagen crosslinking, as first demonstrated by Verzar [30]. This results in a relative stabilization of collagen fibers, resisting degradation as compared to elastic fibers, much poorer in available lysines for GP-mediated crosslinking as compared to collagen. Most lysine residues in tropoelastin, the soluble precursor of elastic fibers, are involved in and modified by the crosslinking process during elastogenesis. This results in progressive rigidification of the vascular wall, as a result of fragmentation of elastic fibers and crosslinking of collagen fibers. The same crosslinking process is occurring in the fine collagen network between heart muscle cells. As an increasing number of these contractile cells are lost with age, the collagen crosslinking will further impede the coordination of rhythmic contractions of the heart muscle. Therefore the Maillard reaction products might well contribute to arrythmias and final heart failure [31].

Collagen crosslinking is a serious contributor to ECM-aging all over the connective tissues of the organism. This is particularly evident for the skin. Its aging is the result of several relatively well-studied mechanisms [32]. Here again the Maillard reaction was shown to play an important role.

Discussion

The Maillard reaction can be considered as ‘illicite’ organic chemistry taking place in the organism, bypassing the gene-controlled metabolic pathways [8]. A link exists, however, between the ‘official’ metabolic reactions and the ‘illicite’ organic chemistry attributed to the Maillard reaction. This link is the receptor mediation of a number of the GP-mediated harmful reactions [33]. Several of these harmful reactions do not need, however, to proceed by receptor-mediated pathways. This is the case for collagen crosslinking as well as for free radical-mediated degradation of biopolymers as shown above. On the contrary, GP-action on cells might well be receptor-mediated. This was proposed for the SA-β-gal expression as a sign of entrance in the senescent state by J. Campisi which might represent an escape mechanism from malignant transformation mediated by anti-oncogens [34]. Cell death, however, as studied here on successive cell passages for cells cultured in plastic wells as in our above-described experiments was of the necrotic type, possibly the result of free radical attack on the cell membrane. Upregulation of elastase-type endopeptidase activity is partially at least receptor-mediated as mentioned before. We could show that the age-dependent uncoupling of the elastin receptor increases elastase and free radical release in presence of elastin peptides [10]. This type of action can result in apoptotic cell death as shown for lymphocytes cultured with elastin peptides [29]. As mononuclear cells in atherosclerotic lesions express the elastin receptor as do also vascular smooth muscle cells [35], the above summarized mechanism may well be involved in the remodeling of the vascular wall during atherogenesis, accelerating the development of atherosclerotic lesions. Our studies conducted at the cellular-molecular level contributed to the description and elucidation of the nature and mechanism of harmful reactions mediated by GPs involved both in age-dependent modifications of tissue structure and function as well as in the onset of age-related diseases, especially cardiovascular pathologies as exemplified by the above discussed role of collagen crosslinking and elastolysis. Some of these processes as, for instance, the loss of elasticity of elastic fibers as a result of Ca and lipid deposition as well as their age-dependent degradation appear to be the result of the structure of elastic fibers in interaction with their environment [31]. In our mind the Maillard reaction is one of the hardly avoidable processes involved in aging and the onset of age-related pathologies. It is, however, only one of these mechanisms, by far not the only one involved in these processes as shown by the steady increase of life expectancy in affluent countries where ‘unhealthy nutrition’ is also widespread. The resulting increase of obesity with its consequences is, however, a process where the Maillard reaction might well play a role in the predicted decrease of life expectancy [36].