Protein Posttranslational Modification (PTM) by Glycation: Role in Lens Aging and Age-related Cataractogenesis

Abstract

Crystallins, the most prevalent lens proteins, have no turnover throughout the entire human lifespan. These long-lived proteins are susceptible to post-synthetic modifications, including oxidation and glycation, which are believed to be some of the primary mechanisms for age-related cataractogenesis. Thanks to high glutathione (GSH) and ascorbic acid (ASA) levels as well as low oxygen content, the human lens is able to maintain its transparency for several decades. Aging accumulates substantial changes in the human lens, including a decreased glutathione concentration, increased reactive oxygen species (ROS) formation, impaired antioxidative defense capacity, and increased redox-active metal ions, which induce glucose and ascorbic acid degradation and protein glycation. The glycated lens crystallins are either prone to UVA mediated free radical production or they attract metal ion binding, which can trigger additional protein oxidation and modification. This vicious cycle is expected to be exacerbated with older age or diabetic conditions. ASA serves as an antioxidant in the human lens under reducing conditions to protect the human lens from damage, but ASA converts to the pro-oxidative role and causes lens protein damage by ascorbylation in high oxidation or enriched redox-active metal ion conditions. This review is dedicated in honor of Dr. Frank Giblin, a great friend and superb scientist, whose pioneering and relentless work over the past 45 years has provided critical insight into lens redox regulation and glutathione homeostasis during aging and cataractogenesis.

Introduction

Cataracts are a common cause of blindness worldwide, particularly in the aged population. More than 70% of Caucasians, 53% of African Americans, and 61% of Hispanic Americans in the US have cataracts by the age of 80. Aging brings substantial changes to the human lens, including increased reactive oxygen species (ROS)(Beebe et al., 2010; Berthoud and Beyer, 2009; Giblin, 2000; Lou, 2003; Truscott, 2005), a decreased antioxidative defense capacity(Wei et al., 2014), disrupted glutathione (GSH) homeostasis(Giblin et al., 1976; Giblin et al., 1985; Sethna et al., 1982), impaired crystallin chaperone function(Andley, 2007; Horwitz, 1992), elevated protein truncation(Schey et al., 2020; Sharma and Santhoshkumar, 2009), and accumulated protein posttranslational modification (PTM), i.e., deamination, deamidation, racemization, thiolation, carbamylation, methylation, phosphorylation, acetylation, and glycation(Gutierrez et al., 2016; Harding, 1991; Lampi et al., 2014; Palsamy et al., 2012; Schaefer et al., 2003; Schey et al., 2021; Tessier et al., 1999; Truscott and Friedrich, 2019). These changes are considered to be the cause of increased lens discoloration, crystallin aggregation, and light scattering, which are the underlying mechanisms of cataractogenesis. Lens protein glycation and accumulative advanced glycation end product (AGE) formation is believed to be one of the major mechanisms in age-related cataractogenesis. Over the years, several excellent reviews have covered the topic of glycation in the eye and eye diseases(Bejarano and Taylor, 2019; Harding, 1991; Kandarakis et al., 2014; Nagaraj et al., 2012; Sharma and Santhoshkumar, 2009; Stitt, 2005). In the present review, we will focus mainly on glycation in lens aging and cataractogenesis. We will predominantly summarize the progress of metabolic pathways of ascorbic acid in the lens that impact the level of individual AGE formation.

Glycation and advanced glycation end products (AGEs)

Glycation is initially defined as a series of non-enzymatic complex reactions between reducing carbohydrates and proteins, lipids, and DNA, also known as the Maillard reaction(Cerami et al., 1979). The definition of glycation has expanded lately into a broader term that also includes non-enzymatic protein/lipids/DNA modification by metabolites, such as methylglyoxal (MGO), glyoxal (GO), as well as ascorbic acid (ASA) metabolites(Linetsky et al., 2008; Rabbani et al., 2016). These reactions proceed over an extended period, resulting in a heterogeneous complex of protein/lipids/DNA covalently bound compounds named advanced glycation end products (AGEs). Levels of most AGEs increase with age and accumulate significantly in tissues with slow or no protein turnover, such as collagen-rich tissues, neuronal tissues, and lens, and are often further exacerbated in diabetes mellitus(Horie et al., 1997; Oudes et al., 1998; Sell et al., 2005; Sell et al., 1996). From a structural viewpoint, glycation causes spontaneous protein damage by changing the protein structure, inducing cross-linking, and modifying protein function(Yan and Harding, 1997; Yim et al., 2001). There is also growing evidence suggesting that glycation and AGE formation not only serve as biomarkers for aging and diseases, but also behave as a pathogenic mechanism in diabetes and diabetic complications, neurodegenerative diseases, obesity, cardiovascular disease, kidney disease, and eye diseases (Rabbani and Thornalley, 2018; Rowan et al., 2018).

Glycation and AGEs in the human lens

In 1981, Monnier and Cerami(Monnier and Cerami, 1981) reported that bovine lens protein, after incubation with glucose or glucose-6-phosphate, generated similar chromophores (~350nm) as those from aged and cataractous human lenses. This ground-breaking study sparked substantial interest in lens glycation research, and 628 publications under “lens glycation” were searchable from PubMed by the end of the year 2020. Early studies determined the lens protein-bound AGEs by either measuring non-tryptophan AGE specific fluorescence, immunohistochemistry using antibodies recognizing pools of AGEs, or enriching glycated proteins through boronic acid affinity chromatography(Araki et al., 1992; Garlick et al., 1984; van Boekel and Hoenders, 1992). Despite heterogeneous AGE mixture and potential interferences by non-AGE contaminants, these studies nevertheless brought valuable information and suggested that lens glycated crystallins were positively correlated with lens aging and further increased in cataractous and diabetic lenses(Kessel et al., 2004; Oimomi et al., 1988). Unequivocal evidence for the occurrence of the Maillard reaction in the lens was achieved with the identification of individual AGE structures , such as N-carboxymethyl-lysine (CML)(Ahmed et al., 1986), pentosidine(Sell and Monnier, 1989) (Nagaraj et al., 1991), methylglyoxal-hydroimidazolones (MG-H1,2,3)(Lo et al., 1994), argpyrimidine (Padayatti et al., 2001) and glucosepane(Biemel et al., 2002). These structures made it possible to develop AGE-specific antibodies, reveal the potential mechanisms of formation, and carry out full chemical synthesis(Nagai et al., 2003; Rosenberg and Clark, 2012; Xue et al., 2014). The precision analysis and quantification of individual AGEs by gas or liquid chromatography-mass spectrometry (GC/MS or LC/MS) technology further opened the field and allowed dissection of complex AGE compositions and pathways in the lens and their role in aging and cataractogenesis(Abraham et al., 1994; Ahmed et al., 1986; Dunn et al., 1989; Patrick et al., 1990; Thornalley et al., 2003). Collectively, close to 30 AGEs have been identified, and we anticipate that even more AGE structures will be identified in the future.

Lens crystallins are lifelong proteins with no turnover. Therefore, they are highly susceptible to glycation and age-related AGE accumulation. Indeed, almost all known AGEs have been found to be present in the human lens (Fig.1). Some of these modifications target the guanidino group of arginine residues while some target the epsilon amino group of lysine residues and others involve both lysine and arginine residues. The lens has very low oxygen content and is therefore considered an anaerobic environment(McNulty et al., 2004). The lens is also an avascular tissue with low glucose content (~2mM)(Tomana et al., 1984). These unique characteristics result in a different metabolism and pathways of AGE formation relative to the aerobic organ environment. In 2010(Fan et al., 2010), we conducted a comprehensive comparative analysis of protein glycation and oxidation between the human lens and skin, anaerobic vs. aerobic tissue. In common, almost all AGEs demonstrate an age-related increase in both human lens and skin. However, despite all measured AGEs being detectable in both tissues, patterns of AGEs present in each tissue were distinct. MG-H1 (500-7000 pmol/mg protein) and CML (100-1500 pmol/mg protein) are the dominant forms of protein damage in aged human lenses, followed by fructose lysine (FL) (200-450pmol/mg protein), CEL (10-700 pmol/mg protein), glucosepane (5-500 pmol/mg protein), and K2P (5-400 pmol/mg protein) (Fig.2). The human lens has 5 times less FL and glucosepane modification than skin collagen, which reflects the low glucose content in the human lens because both FL and glucosepane are uniquely derived from glucose. Nevertheless, due to long-lived nature of lens crystallins, glucose still induces substantial protein damage in aged lenses. In 2015, Glomb’s team(Smuda et al., 2015) surveyed 19 AGEs in healthy and cataractous human lenses and reached a similar conclusion in which CML (400-1500 pmol/mg protein) and MG-H1 (200-1900 pmol/mg protein) are the two dominant AGEs in the aged and cataractous human lenses. They also found two AGEs, the C4 and C5 amides that are unique glycation products derived from ascorbic acid. Also, AGEs have been detected from the human lens capsule(Raghavan et al., 2016).

Figure 1.

Identified AGEs that are present in human lenses.

Figure 2.

Comparison of levels of AGEs in human lenses and skin collagen. The data were expressed as mean ± SD. Inset: AGEs with levels below 600 pmol/mg protein. Mean levels with a different letter superscript are significant (p<0.05) as determined by the Mann-Whitney rank test.

Age-associated accumulation of AGEs is believed to play a causative role in age-related cataract formation. First, glycation and AGE formation can cause lens crystallin crosslinks, aggregation, and high molecular weight (HMW) protein formation. Numerous in vitro studies involving incubation of lens protein extract or isolated crystallins with glucose or other glycation reagents, such as methylglyoxal (MGO), demonstrate a remarkable increase in crystallin aggregation(Biemel et al., 2002; Chellan and Nagaraj, 2001; Fuentealba et al., 2009; Swamy et al., 1993). Van Boekel et al. (van Boekel and Hoenders, 1991) showed that the incubation of calf lens crystallins with glucose-6-phosphate increased the mean molecular weight of all crystallins based on high-pressure gel permeation chromatography, SDS-PAGE, and analytical ultracentrifugation analysis. Prabhakaram et al. (Prabhakaram and Ortwerth, 1992) tested individual calf lens crystallins incubation with ascorbic acid for 3 weeks and found a substantial amount of HMW crosslink from both alpha and beta crystallin. Karumanchi et al. (Karumanchi et al., 2015) found that alpha A crystallin (αA), when incubated with 10μM MGO for 9hr at 25°C, increased the particle radius from 18nm (native) to 350nm. Lens crystallins can also form covalently bound crosslinks with other lens proteins. For example, a covalent crosslink between lens membrane protein aquaporin 0 (AQP0) and αA crystallin was formed when αA crystallin was incubated with bovine lens membranes in the presence of dehydroascorbic acid (DHA) or threose (Prabhakaram et al., 1996). Notably, such crystallin and membrane crosslinks were found highly present in aged human lenses and were significantly exacerbated in cataractous and diabetic lenses(Prabhakaram et al., 1996). In vivo data very well echo these in vitro findings. Das et al.(Das et al., 1998) showed that the water-insoluble fraction (WIF) had much higher amounts of AGEs than the water-soluble fraction (WSF) and was also positively correlated with lens nuclear discoloration. A later study by Zarina et al. (Zarina et al., 2000) compared water-soluble (WSF), urea-soluble (USF), alkali-soluble (ASF), sonicated (SF), sonicated-insoluble (SI), and membrane (MF) fractions of human lens protein extracts from both human senile and diabetic cataractous lenses and found that AGEs accumulated mostly in the high molecular aggregates in all fractions. In a similar study by Pokupec et al. (Pokupec et al., 2003) in which AGEs content was surveyed in nondiabetic and diabetic human lenses, and found that higher levels of AGEs were significantly associated with HMW proteins. These findings suggest that lens crystallin glycation and AGE formation contribute to crystallin aggregation and HMW formation.

Glycation often induces lens protein structural changes that result in protein instability and dysfunction(Liang and Fu, 2001; Yousefi et al., 2016). An early study by Luthra et al. (Luthra and Balasubramanian, 1993) indicates that long-term (8 weeks) modification of γ-crystallin by fructose can induce partial crystallin unfolding and protein instability that make them prone to thermodynamic and chemical mediated denaturation. Interestingly, in the very same study, the opposite behavior was seen in α-crystallin incubation. A recent report (Yousefi et al., 2016) also showed a partial resistance of α-crystallin structure alteration after 4 weeks of glucose incubation. The α-crystallin’s chaperone function and relative low glycation reactivity of glucose and fructose may be the reasons behind these findings. In contrast, when α-crystallin was modified by MGO, Mukhopadhyay et al. (Mukhopadhyay et al., 2010) reported drastic reduction of denaturation temperature and chaperone function of α-crystallin after 5 days of MGO modification. Similarly, Kumar et al. (Kumar et al., 2004b) showed that MGO modification increased α-crystallin instability making it prone to degradation due to partial unfolding. Studies also demonstrated that glycation could trigger simultaneous inactivation of several key intracellular enzymes, such as catalase, superoxide dismutase(Yan and Harding, 1997), glucose-6-phosphate dehydrogenase (G6PDH)(Zhao et al., 1998), malate dehydrogenase (MDH)(Heath et al., 1996), and NADP(+)- dependent isocitrate dehydrogenase (ICDH)(Kil et al., 2004).

α-crystallin acts as a molecular chaperone and is considered to play a pivotal role in maintaining lens transparency(Derham and Harding, 1999; Horwitz, 1992). An early study by van Boekel et al.(van Boekel et al., 1996) suggests that late but not early glycation decreases α-crystallin chaperone function. Kumar et al. (Kumar et al., 2007) tested the impact of various glycating reagents on α-crystallin chaperone activity and found that glucose, fructose, and glucose-6-phosphate mediated glycation could all impair α-crystallin’s chaperone-like function. Ortwerth group (Bhattacharyya et al., 2007) performed an elegant study through chemically conjugating N-(2-bromoethyl)-3-oxidopyrimidinium hydrobromide to cysteine residue from K90C mutated αB to create an OP-lysine modified αB crystallin and found that a single lysine site modification significantly decreased intrinsic tryptophan fluorescence and αB chaperone activity. Furthermore, by blocking glycation at lysine residue 11, 78, and 16 of αA and lysine residue 90, 92, and 166 of αB, crystallin chaperone activity that was affected by glycation could be restored(Abraham et al., 2008). Interestingly, α-crystallin chaperone activity was reduced only after being extensively modified with MGO (~100mM) (Derham and Harding, 2002). Nagaraj’s group showed that MGO modification significantly enhanced α-crystallin activity(Kanade et al., 2009; Nagaraj et al., 2003; Puttaiah et al., 2007). The underlying mechanisms of augmented α-crystallin chaperone function by MGO remains unknown, but the study suggested that the increased α-crystallin chaperone-like activity of MGO modification is not associated with increased surface hydrophobicity(Kumar et al., 2004a). Interestingly, Nagaraj’s group recently suggested that α-crystallin chaperone function mediated chaperone-client complexes may actually facilitate inter-protein cross-link due to their close proximity(Nandi et al., 2020a; Nandi et al., 2020b).

Glucose vs. Ascorbic Acid

As mentioned earlier, the lens has low glucose content and therefore contains low levels of fructose-lysine (FL) in both young and aged human lenses. On the other hand, the human lens has high levels (3-5mM) of ascorbic acid (ASA), also known as Vitamin C(Fan et al., 2006). ASA is not a very stable compound and can readily lose two electrons while being oxidized(Simpson and Ortwerth, 2000). Several ascorbic acid oxidation intermediates are highly reactive and can cause protein damage by forming covalent protein-bound AGEs, often described as ascorbylation. Ortwerth’s group (Lee et al., 1998) quantitatively compared the glucose, fructose, and ASA reactivity of their in vitro incubation with bovine lens protein using universal ¹⁴C radioactive labeled glycating reagents. Astonishingly, they found that ASA was 18 times more reactive than glucose in its incorporation into lens protein, while ASA was 70 and 100 times more active than glucose and fructose in protein crosslinking, respectively. However, the ascorbylation mediated protein damage is derived from ASA oxidation and degradation. In its reduced form, ASA is not likely to cause protein damage via protein modification, i.e., glycation. The high concentration of ASA in the healthy human lens, particularly the young human lens, generally acts as an antioxidant to protect the lens from damage(Hegde and Varma, 2004; Lim et al., 2020; Valero et al., 2002). For example, ASA is a highly effective UV filter reagent protecting the lens from UV-mediated damage. During lens aging, impaired redox homeostasis and accumulated ROS levels will likely induce ASA oxidation and protein ascorbylation. Also, despite low glycating activity by glucose and its low presence in the lens, glucose through glycolysis and degradation can produce highly reactive metabolites (e.g. MGO) that can cause protein damage and crosslink via simultaneous AGEs formation. On top of this, diabetes mellitus often markedly increases lens glucose levels(Bron et al., 1993; Bursell et al., 1989).

Ascorbic acid degradation

Although ASA is the dominant form present in tissues, including the lens, approximately 5% dehydroascorbic acid (DHA) equilibrium can always be detected in tissue homogenate. Glutathione is believed to play a crucial role in maintaining ascorbic acid in its reduced form. When there is an excessive amount of DHA, it can be converted to ASA by glutathione-mediated reduction(Rose and Bode, 1992). However, DHA has a short half-life time (~2min) and can be readily hydrolyzed to 2,3-diketogulonic acid (2,3-DKG), an irreversible ASA metabolite. 2,3-DKG has a relatively longer half life time (~2.5hr)(Simpson and Ortwerth, 2000). Using the ¹³C-NMR method, Simpson et al.(Simpson and Ortwerth, 2000) determined ASA degradation products under physiological conditions. They found that L-erythrulose (ERU) was the primary product (over 97%) from the non-oxidative degradation of DHA and 2,3-DKG. ERU formation is believed to occur through carbon 2 and 3 cleavage of 2,3-DKG. In oxidative degradation of DHA and 2,3-DKG, oxalate and CO₂ are the main products. In 2013, Smuda et al.,(Smuda and Glomb, 2013) determined the ASA Maillard reaction degradation pathway by incubating ASA with alpha amine protected lysine (N_α-Boc-lysine) and analyzed the products using LC/MS/MS. Over 15 degradation products were identified in this study, and importantly, ERU was also predicted to be the major non-oxidative degradation product. In 2011, Nemet et al.(Nemet and Monnier, 2011) surveyed ASA degradation in the human lens using an elegant o-phenylenediamine trapping approach. They were able to detect five ASA degradation products, DHA, 2,3-DKG, 3-deoxythreosone, xylosone, and threosone, from both a protein-free fraction and a water-soluble fraction (WSP). Importantly, 3-deoxythreosone was the major degradation product in both factions, which confirms the in vitro findings that ERU is the major degradation product resulting from ASA’s non-oxidative degradation (Fig.3).

Figure 3.

Non-oxidative and oxidative ASA degradation pathways.

The human lens is constantly exposed to sunlight, including UVA/UVB rays. Linetsky et al. (Linetsky et al., 1996) demonstrated that UVA irradiation of aged human lens water-soluble protein could generate a significant amount of superoxide anion and hydrogen peroxide. Conversely, after incubation with ASA, lens protein significantly increases UVA-mediated superoxide anion formation. Thus, a free radical and ASA oxidation vicious cycle is formed in aged human lenses(Linetsky and Ortwerth, 1996). Remarkably, UVA can trigger ASA oxidation even in the presence of high levels of glutathione (GSH)(Ortwerth et al., 1998), suggesting that the initial ASA oxidation product, i.e., DHA, is further oxidized by singlet oxygen to an irreversible degradation product, i.e., 2,3-DKG. A study also indicates that UVA can induce ASA oxidation in the absence of oxygen(Ortwerth et al., 2003). More recently, Linetsky et al.(Linetsky et al., 2014) showed that kynurenines could facilitate UVA mediated ASA oxidation and protein modification (Fig.3).

Ascorbic acid mediated glycation in lens aging and cataractogenesis

In 1985, Bensch et al.(Bensch et al., 1985) found that ASA could trigger lenticular protein “browning,” and the formation of yellow and brown condensation was correlated with ASA incorporation determined by C¹⁴ labeled ASA. In 1998, Ortwerth’s group (Ortwerth et al., 1988; Ortwerth and Olesen, 1988) demonstrated that in vitro incubation of lenticular protein with ASA but not glucose produced crystallin crosslinks and high molecular weight (HMW) protein suggesting that ASA induces Maillard reactions in the human lens. Nagaraj et al. (Nagaraj et al., 1991) showed that ASA and its oxidation products contributed to the pentosidine formation during lens aging and cataractogenesis. Vesperlysine A/LM-1, an AGE that can be easily produced from protein incubation with ribose, ASA, DHA, and 2,3-DKG, was found in human lenses and positively associated with aging(Tessier et al., 1999). Aging lenses manifest discoloration with accumulated fluorophores and chromophores. Interestingly, chromophores produced from calf lens protein and ASA incubation were found to closely mimic those isolated from human cataractous lenses(Cheng et al., 2001). In 2006, Ortwerth’s group(Cheng et al., 2006) performed a sophisticated study by comparing water-insoluble protein (WIS) modifications between aged healthy human lenses, early-stage brunescent cataractous lenses, and calf lens protein reacted with or without 20mM ASA via two-dimensional LC/MS mapping technique. Remarkably similar modifications were seen in aged and early-stage brunescent cataractous lenses compared to ASA modified lens proteins. Intriguingly, approximately 100 modification sites were similar between groups, 40 major modification spots were analyzed, and 90% of them were identical.

To further support studies on the role of ascorbylation in the aging lens, we took advantage of the fact that mouse lenses have barely detectable ASA and created a humanized transgenic mouse model by specifically overexpressing the sodium-dependent Vitamin C transporter 2 (hSVCT2) throughout the entire lens. The hSVCT2 transgenic mice lenses expressed similar ASA levels as in human lenses and demonstrated increased levels of AGEs compared to wild-type mice lenses and manifested yellow discoloration at 12 months of age which is similar to human lenses aged in their 60s (Fig.4). Topical treatment with L-arginine suppressed protein-linked fluorescence, CML, CEL and glucosepane but not oxoaldehyde hydroimidazolones (Fan and Monnier, 2008) (Fan et al., 2011).

Figure 4.

12 months old hSVCT2 lens specific transgenic mouse lenses manifest a similar lens discoloration as 67 years human lens.

In 2010(Rautiainen et al., 2010), an 8.2 year follow-up study compared the incidence of age-related cataract extraction in women between vitamin C supplement users and non-vitamin C users. Surprisingly, for women aged 65 or older, vitamin C supplement takers had a 38% higher risk of developing cataracts and a 56% higher cataract risk for vitamin C supplement takers under hormone replacement therapy. These findings strongly suggest that ASA oxidation and ascorbylation may play a crucial role in lens aging and cataractogenesis.

The biochemical origin of MG-H1 in the lens

Oxoaldehyde stress is recognized as a significant threat to health in various metabolic conditions and diseases of accelerated aging, including age-related cataracts (Rabbani and Thornalley, 2015; Saeed et al., 2020). The common oxoaldehydes present in the lens are methylglyoxal (MGO) and glyoxal (GO). Approximately 1.8μM MGO was reported in the human lens by enzymatic based assay(Haik et al., 1994), although the newly developed stable isotopic dilution LC/MS method suggests that this level might be overestimated (Rabbani and Thornalley, 2014). No actual lens glyoxal levels were reported, but it is expected to be very low due to minimal carboxymethyl arginine (CMA) modification, a sole modification by GO, was detected in the human lenses(Smuda et al., 2015). MGO is a metabolite resulting from the triose isomerase step of glycolysis and is present as a minor source in Maillard catalyzed degradation of reducing sugars. MGO is highly reactive and can cause protein damage by directly forming AGEs. Several AGEs, such as methylglyoxal lysine dimer (MOLD), carboxymethyl-lysine (CML), argpyrimidine, methylglyoxal-derived lysine-arginine crosslink (MODIC), and the methylglyoxal-derived hydroimidazolones (MG-H1, MG-H2, and MG-H3) are generally thought to be derived from MGO. Among these AGEs, the most abundant MGO derived AGEs is MG-H1. MG-H1 is believed to be the most prevalent cellular AGE in animal tissues. It has been suggested that at least one MG-H1 adduct is present in 3-13% of protein, corresponding to about 1–2% of total arginine residues of proteins found in mammalian tissue, plasma, and extracellular matrix(Rabbani and Thornalley, 2012). In aged human lenses, for example, Thornalley et al. found levels of MG-H1 up to 14 nmol/mg protein(Ahmed et al., 2003), while our comparative studies confirmed that MG-H1 is the most abundant AGEs in aged human lenses(Fan et al., 2010). In the lens, such levels are significant as they may affect crystallin conformation and chaperone function(Kumar et al., 2007; Mukhopadhyay et al., 2010).

Our recent study suggests that ASA catabolites might be an essential source for MG-H1 formation(Fan et al., 2020). The carbons 4-5-6 of ASA are incorporated into the MG-H1 structure, suggesting a cleavage between carbon 3 and 4 of ASA that likely resulted from erythrulose. In vitro incubation of young human lens protein extract with physiological concentration (3mM) of ASA readily produces a significant amount of MG-H1 (~2500 pmol/mg protein), but not from 30mM glucose incubation. The amount of MG-H1 produced from 3mM ASA is equivalent to 8μM MGO-mediated modification, suggesting that ASA contributes significantly to lens protein damage by forming MG-H1 modification. The prominent role of ASA in MG-H1 formation is also seen in the lens-specific SVCT2 transgenic mouse (hSVCT2), showing several folds elevated MG-H1 levels. Interestingly, ASA-mediated MG-H1 formation is also seen in brain tissue, another ASA enriched organ. Therefore, oxoaldehyde stress might originate from ascorbic acid rather than glucose in ascorbic acid-rich cells as in neurons and lens fiber cells.

Nevertheless, the precise biochemical origin of MG-H1 in the lens may be unrelated to methylglyoxal per se. Indeed, when hSVCT2 mouse lenses were topically treated with L-arginine drops, several AGEs other than MG-H1 formation were suppressed (Fan et all 2008). Thus, more likely, the ascorbate-derived glyceraldehyde is the immediate MG-H1 precursor by binding directly to the guanidino group and closing the hydroimidazolone ring under mild oxidation.

Aging, redox active metal ions, ASA oxidation, and glycation

Metal ions, such as transition metal ions, can have a significant impact on protein glycation. For example, surveying the glycated plasma hemoglobin in diabetic patients reveals that it is positively correlated with plasma copper levels(Viktorinova et al., 2009). Transition metal ions have also been found facilitating glucose autoxidation to glyoxal (GO), a highly reactive glycating metabolite(Wells-Knecht et al., 1995). Copper ion-mediated ASA oxidation could also produce GO and increase CML formation(Shangari et al., 2007). In 2009, we teamed up with Dr. Frank Giblin and found that the lysine oxidation product adipic semialdehyde (allysine), and its oxidation end product 2-aminoadipic acid, significantly increased with age and strongly correlated with cataract grade. Importantly, such oxidation was best explained by metal-catalyzed oxidation likely facilitated by ASA oxidation(Fan et al., 2009).

Aging brings substantial changes to lens metal ion homeostasis. A number of studies targeting lens metal ions composition yielded mixed results. Several studies report that some metal ions, such as Fe, Cu, Zn, Cd, Ca, Se, and Pb are increased in cataractous human lenses compared to healthy lenses(Cekic, 1998; Davidson et al., 1998; Erie et al., 2005; Gozzelino and Arosio, 2016; Hou and Hou, 1996; Srivastava et al., 1992). At the same time, other studies suggest no significant changes in the level of Cu, Fe, and no detectable Pb in human cataractous vs. healthy lenses(Cekic, 1998; Langford-Smith et al., 2016; Rajkumar et al., 2013). Smoking has also been involved in increasing lens Fe and Ca content(Avunduk et al., 1997). Metal ion homeostasis is tightly regulated, e.g., disruption of iron homeostasis has been found to be directly involved in various diseases(Davidson et al., 1998; Gozzelino and Arosio, 2016). Garner et al.(Garner et al., 2000b) found that the level of redox-active iron, also known as labile iron pool (LIP), was significantly higher in cataractous lenses compared to non-cataractous lenses. LIP has been revealed to promote ROS formation, such as hydroxyl free radical formation via Fenton-like reactions. A considerable amount of hydroxy radical attack on lens proteins has been found from nuclear cataractous lenses(Fu et al., 1998; Garner et al., 2000a). The accumulative redox-active iron can switch ASA’s antioxidative role to its pro-oxidative role(Fan et al., 2020). ASA can readily reduce ferric (Fe₃⁺) ions to ferrous (Fe₂⁺) irons while oxidizing itself into dehydroascorbic acid (DHA) (see reaction 1). The ferrous (Fe₂⁺) iron can then convert hydrogen peroxide into hydroxy free radical (HO^•) (see reaction 2). Again, a vicious cycle is formed in aged and cataractous human lenses that will favor ASA oxidation and protein modification by glycation.

$$\text{ASA+F}{\text{e}}_3^{+}=\text{F}{e_2}^{+}+\text{DHA}$$ (1)

$$\text{F}{\text{e}}_2^{+}+{\text{H}}_2{\text{O}}_2=\text{F}{\text{e}}_3^{+}+{\text{HO}}^{•}+{\text{OH}}^{-}$$ (2)

Lens protein may have ways to protect transition metal-mediated oxidation. Ortwerth et al. (Ortwerth and James, 1999) showed that lens protein could tightly bind copper ions (Cu²⁺), blocking ASA oxidation and reactive oxygen species (ROS) formation. However, with the formation of AGEs, such as carboxymethyllysine (CML), Saxena et al. (Saxena et al., 2000) found that glycated protein with AGEs such as CML could bind redox-active copper ions and induce oxidation, e.g., CML-rich proteins oxidized approximately 4 times faster than similar protein from young and aged healthy human lenses.

Concluding remarks and perspectives

The lens is a unique organ with low oxygen, unusually high glutathione and ascorbic acid levels, and dense but orderly packed long-lived proteins. Age-related accumulation of protein damage by oxidation and posttranslational modification (PTM) will likely result in protein aggregation and a higher risk of age-related cataract formation. Compelling evidence from over 40 years of research indicates that lens protein PTM by the Maillard Reaction may play a pathogenic role in age-related cataractogenesis. Accumulating AGEs can be derived, not only from glucose but from various metabolites, and ascorbic acid degradation products. Although significant progress has been achieved in the past 40 years, a number of questions remain to be addressed. First, the question of site specificity and glycation hot spots of lysine, arginine, and histidine residue needs to be better understood. Second, more studies are needed to understand the impact of glycation on crystallin structure and physiological function. For example, does MG-H1 induce α-, β-, and γ-crystallin partial unfolding? Does MG-H1 affect crystallin-crystallin interaction? Does MG-H1 affect α-crystallin chaperone function? Third, more studies are needed to clarify whether accumulating AGEs impact the signaling cascades and therefore lens epithelial cell proliferation, fiber cell differentiation, and maturation. Fourth, additional studies are required to understand the relationship between lens protein glycation and lens metabolism. In that regard, how metabolic changes during aging impact lens protein glycation is mostly unknown. For example, the impact of the profound age-related loss of glutathione on glucose and ascorbic acid degradation is still poorly understood. In many ways, these questions grew directly out of the pioneering work of Dr. Frank Giblin on lens aging and redox regulation, particularly the tight relationship between glutathione biology and cataract development.

Highlights.

• Aging is the major risk factor for cataracts.

• Age-related accumulation of PTM by glycation is one of the primary mechanisms in age-related cataractogenesis.

• Ascorbic acid (Vitamin C) oxidation and degradation can trigger lens protein damage by glycation