Anaerobic vs aerobic pathways of carbonyl and oxidant stress in human lens and skin during aging and in diabetes: A comparative analysis

Abstract

The effects of anaerobic (lens) vs aerobic (skin) environment on carbonyl and oxidant stress are compared using de novo and existing data on advanced glycation and oxidation products in human crystallins and collagen. Almost all modifications increase with age. Methylglyoxal hydroimidazolones, carboxymethyllysine, and carboxyethyllysine are severalfold higher in lens than in skin and markedly increase upon incubation of lens crystallins with 5mM ascorbic acid. In contrast, fructose-lysine, glucosepane crosslinks, glyoxal hydroimidazolones, metal-catalyzed oxidation (allysine), and H(2)O(2)-dependent modifications (2-aminoapidic acid and methionine sulfoxide) are markedly elevated in skin, but relatively suppressed in the aging lens. In both tissues ornithine is the dominant modification, implicating arginine residues as the principal target of the Maillard reaction in vivo. Diabetes (here mostly type 2 studied) increases significantly fructose-lysine and glucosepane in both tissues (P<0.001) but has surprisingly little effect on the absolute level of most other advanced glycation end products. However, diabetes strengthens the Spearman correlation coefficients for age-related accumulation of hydrogen peroxide-mediated modifications in the lens. Overall, the data suggest that oxoaldehyde stress involving methylglyoxal from either glucose or ascorbate is predominant in the aging noncataractous lens, whereas aging skin collagen undergoes combined attack by nonoxidative glucose-mediated modifications, as well as those from metal-catalyzed oxidation and H(2)O(2).

Fig. 1

Overview scheme of the Maillard reaction in vivo. Selected pathways linking potential carbonyl sources in lens vs. skin with those advanced glycation endproducts that are measured in this study.

Fig. 2

Top panels: Age-related changes in levels of fructose-lysine and glucosepane in insoluble lens crystallins and skin collagen-rich fraction. Regression lines are for fructose-lysine in skin, y = 1562 + 7x , n=12, r=0.46, P=0.11 (NS); in lens, y = 265 + 2x, n= 17, r=0.46, P=0.064 (NS); for glucosepane in skin, y = 20 + 0.034x, n=14, r=0.82, P<0.0001; in lens, y = 44 + 4x, n=25, r=0.72, P< 0.0001. The accumulation rates of fructose-lysine between skin vs. lens are statistically the same (P>0.2, NS). For glucosepane, the accumulation rate is significantly greater for skin vs. lens (P<0.0001). Middle and lowers panels: Effects of diabetes (mostly type 2 studied) on fructose-lysine and glucosepane in human lens crystallins and skin. The 95% prediction intervals are shown for nondiabetic individuals. Regression analyses comparing levels vs. age and diabetes shows that diabetes (lens, P<0.0001, skin P=0.006) is significant for both fructose-lysine and glucosepane. Age is significant for glucosepane (P<0.0001) but not fructose-lysine (P>0.06). Each individual is indicated by a symbol: ●, nondiabetic lens; ○, nondiabetic skin; ▲, diabetic. Regression lines for lens and skin data are shown as solid and dash lines, respectively. NS: nonsignificant.

Fig. 3

Top panels: Age-related changes in carboxymethyl-lysine (CML), glyoxal hydroimidazolone (G-H1) and pentosidine in insoluble lens crystallins and skin collagenrich fraction. Regression lines are (see Fig. 2): CML, skin, y = 5x − 84, n=13, r=0.78, P=0.001; lens, y = 152 + 10x, n= 25, r=0.54, P=0.004; G-H1, skin, y = 6x − 137, n=13, r=0.86, P<0.0001; lens, y = 0.6 + 0.02x, n=24, r=0.15, P=0.47 (NS); pentosidine, skin, y = 0.14 + 0.33x, n=15, r=0.73, P=0.002; lens, y = 0.16x – 3.3, n=22, r=0.65, P=0.001 with 3 outliers. CML: levels are significantly higher in lens vs. skin (P<0.0001) but the accumulation rates are nonsignificantly different (P>0.25). GH-1: The accumulate rate and levels are significantly higher in skin vs. lens (P<0.0001). Pentosidine: Levels are significantly more elevated (P<0.0001) for skin vs. lens, but the accumulation rates are nonsignificantly different (P>0.05). Middle and lower panels: the effect of diabetes is nonsignificant (P>0.05) for all three parameters for both tissues.

Fig. 4

Top panels: Age-related changes in levels of methylglyoxal hydroimidazolone (MG-H1) and carboxyethyl-lysine (CEL) in insoluble lens crystallins and skin collagen-rich fraction. Top panel: Regression lines are (see Fig. 2): MG-H1, skin, y = 5x − 66, n= 14, r=0.64, P=0.014; lens, y = 69x − 181, n= 25, r=0.88, P<0.0001; CEL, skin, y = 2x − 16, n=14, r=0.59, P=0.027; lens, y = 254 + 3x, n=25, r=0.34, P=0.099 (NS). The accumulation rate is significantly (P<0.0001) greater for lens vs. skin for MG-H1, but not CEL (P>0.6). Levels are significantly higher in lens vs. skin for both parameters (P<0.0001). Middle and lower panels: the effect of diabetes is nonsignificant (P>0.1) in levels of MG-H1 and CEL for both tissues.

Fig. 5

Formation of AGEs in bovine lens crystallin homogenate incubated with 5 mM ascorbic acid or DHA for 7 days in 100 mM phosphate buffer containing 1 mM DTPA chelating agent. All values are elevated in presence of ASA and DHA compared to control (p<0.001)

Fig. 6

Age-related accumulation rates of allysine, 2-aminoadipic acid, pentosidine and ornithine in skin vs. lens. The regression lines and 95% confidence intervals of prediction are shown for both skin (dash lines) and lens (solid lines with shading). Top panels: The data for skin (nonsepsis) and lens is replotted from [39] and Fan [22], respectively. Regression lines are (see Fig. 2): allysine, skin, y = 71 – 0.02x, P=0.94 (NS); lens, y = 5 + 0.7x, P<0.0001; 2-aminoadipic acid, skin, y = 78 + 3x, P<0.0001; lens, y = 28 + 0.6x, P=0.042. The accumulation rate is significantly greater in skin vs. lens for 2-aminoadipic acid (P<0.0001), but not allysine (P>0.07). The accumulation rate is higher in skin vs. lens for both allysine (P<0.001) and 2-aminoadipic acid (P<0.0001). Lower panels: Regression lines for skin pentosidine see Fig. 3. ornithine, skin, y = 1281 + 32x, P<0.0001; lens, y = 3445 + 67x, P=0.033. For ornithine, the levels are significantly more elevated for lens vs. skin (P<0.0001), but the accumulation rates are not different (P>0.11).

Fig. 7

Quantitative comparisons of mean levels ± SD for AGEs and methionine oxidation levels reached at age ≥ 65 yrs for insoluble lens crystallins and skin collagen-rich fraction. Levels shown for argpyrimidine were reproduced from non-cataractous lens samples of Padayatti et al. [51] while methionine sulfoxide levels for lens vs. collagen were from Sochaski et al. [34]. Mean levels with a different letter superscript are significantly different (P< 0.05) as determined by the Mann-Whitney rank test.