Autophagic receptor p62 protects against glycation‐derived toxicity and enhances viability

Abstract Diabetes and metabolic syndrome are associated with the typical American high glycemia diet and result in accumulation of high levels of advanced glycation end products (AGEs), particularly upon aging. AGEs form when sugars or their metabolites react with proteins. Associated with a myriad of age‐related diseases, AGEs accumulate in many tissues and are cytotoxic. To date, efforts to limit glycation pharmacologically have failed in human trials. Thus, it is crucial to identify systems that remove AGEs, but such research is scanty. Here, we determined if and how AGEs might be cleared by autophagy. Our in vivo mouse and C. elegans models, in which we altered proteolysis or glycative burden, as well as experiments in five types of cells, revealed more than six criteria indicating that p62‐dependent autophagy is a conserved pathway that plays a critical role in the removal of AGEs. Activation of autophagic removal of AGEs requires p62, and blocking this pathway results in accumulation of AGEs and compromised viability. Deficiency of p62 accelerates accumulation of AGEs in soluble and insoluble fractions. p62 itself is subject to glycative inactivation and accumulates as high mass species. Accumulation of p62 in retinal pigment epithelium is reversed by switching to a lower glycemia diet. Since diminution of glycative damage is associated with reduced risk for age‐related diseases, including age‐related macular degeneration, cardiovascular disease, diabetes, Alzheimer's, and Parkinson's, discovery of methods to limit AGEs or enhance p62‐dependent autophagy offers novel potential therapeutic targets to treat AGEs‐related pathologies.


| INTRODUC TI ON
Americans consume very high glycemic diets, and the trend toward consuming these diets is increasing throughout the world.
Associated with consumption of such high glycemic diets are markedly increased risks for many major age-related debilities including cardiovascular disease (CVD), diabetes, age-related macular degeneration (AMD), some forms of cataract, and neurodegenerative diseases such as Alzheimer's disease and Parkinson's disease (Bejarano & Taylor, 2019;Chaudhuri et al., 2018;Moldogazieva & Mokhosoev, 2019;Vicente Miranda et al., 2016). Alarmingly, the increase in risk for disease due to consuming high glycemic diets is comparable to the risk incurred by smoking. In contrast, consuming lower glycemic diets is associated with slower progression of some of these diseases. These data suggest that switching to lower glycemic diets can reduce the risk of developing several severe medical conditions, bringing tremendous personal and public health benefits. What might be mechanisms for the salutary effect of lower glycemic diets?
Protein glycation results from the non-enzymatic chemical reaction of sugars with proteins. Initial steps are called the Amadori and Maillard reactions. Metabolic products of sugar, some oxidized, such as methylglyoxal (MGO) are primary biological glycating agents. The products can progress through a myriad of rearrangements and additional reactions. Collectively, these are called advanced glycation end products (AGEs). The excess glucose and its metabolic products that result from a high glycemia diet, or diabetes, have been shown to induce and accelerate glycative stress. Even in nondiabetics, AGEs accumulate with accelerating rates upon aging in most tissues, in pathologies such as cataracts and AMD and are cytotoxic (Kazi et al., 2017;Rabbani & Thornalley, 2015;Uchiki et al., 2012;Weikel et al., 2012). We and others have observed elevated levels of AGEs in tissues, including liver, brain, retina, heart, collagen, of nondiabetic animals that consumed high glycemic diets or were aged (Uchiki et al., 2012;Weikel et al., 2012). In contrast, diminishing the level of AGEs has been proven to prolong lifespan in model organisms (Kazi et al., 2017). Efficient removal of AGEs is especially relevant in highly differentiated tissues such as the retina, lens or brain. In such organs, glycation damage cannot be diluted by cellular division and is indicative of disease (Chaudhuri et al., 2018).
The mechanisms by which AGEs damage the individual are poorly understood. It has been proposed that AGEs threaten cellular homeostasis by compromising the function of critical biomolecules, forming dysfunctional toxic aggregates, and recruiting and/or inactivating other essential proteins. Collectively, these insults lead to aberrant metabolism and cellular vulnerability (Rabbani & Thornalley, 2015).
The deposition of AGEs can be limited by detoxification of reactive AGEs precursors such as MGO via the glyoxalase system (Morcos et al., 2008). However, once AGEs are formed, there are no enzymes that specifically remove the added sugar or sugar derivatives from proteins. Two major proteolytic pathways are proposed to contribute to the AGEs clearance: the ubiquitin-proteasome system (UPS) and autophagy (Takahashi et al., 2017;Taylor, 2012). The UPS operates mainly on soluble substrates and uses the proteasome for degradation, whereas autophagy targets cargoes to the lysosomal compartment for degradation. Autophagy is the major degradative route for the clearance of cytosolic, aggregated, or insoluble proteins and organelles that cannot pass through the proteasome.
The UPS and autophagy cooperate functionally, and the deficiency of one of these pathways can trigger the upregulation of the other route (Gavilan et al., 2015;Ji & Kwon, 2017). The function of these degradative pathways declines with age, contributing to the intracellular accumulation of proteinaceous aggregates and dysfunctional organelles in aged tissues (Mizushima et al., 2008). It is presently unknown 1) if the clearance of AGEs is impacted by crosstalk between the UPS and autophagy, 2) if different pools of AGEs are differentially targeted to each pathway, and 3) if upregulating proteolytic pathways increases clearance of AGEs to benefit cell and organismic viability in the face of glycative stress.
During the autophagic process, damaged proteins/organelles or aggregates are tagged with ubiquitin and sequestered into double-membrane structures called autophagosomes. The completion of this process requires the collaboration of a set of autophagic proteins including structural elements involved in the biogenesis of the autophagosome. These include the microtubule-associated protein chain 3 (LC3, a mammalian homologue of yeast Atg8) and receptors that target ubiquitinated cargo to the autophagic compartment. The best studied of these autophagic receptors is p62/SQSTM1/A170/ZIP (hereafter called p62). p62 facilitates selective autophagic clearance as high mass species. Accumulation of p62 in retinal pigment epithelium is reversed by switching to a lower glycemia diet. Since diminution of glycative damage is associated with reduced risk for age-related diseases, including age-related macular degeneration, cardiovascular disease, diabetes, Alzheimer's, and Parkinson's, discovery of methods to limit AGEs or enhance p62-dependent autophagy offers novel potential therapeutic targets to treat AGEs-related pathologies.
Mature autophagosomes fuse with lysosomes that provide digestive enzymes, and both LC3 and p62 are degraded by lysosomal proteases along with the cargo.
Given the apparent deleterious impact of AGEs, there is a need to develop strategies to counteract their accumulation and the disease-related sequelae. There is limited mechanistic information about the targeting of AGEs for degradation, and a lack of understanding about the role of autophagy in this process limits our ability to formulate therapeutic strategies to reduce the risk for AGEs-related diseases. In this study, we explored the contribution of p62-dependent autophagy to the clearance of AGEs. We show for the first time a protective role for p62 against glycation-derived toxicity and identify this autophagic receptor as a novel potential therapeutic target to treat AGEs-related pathologies.

| p62-dependent autophagy plays a role in the clearance of endogenous AGEs
Given that the UPS and autophagy are functionally coupled and are proposed to participate in the removal of AGEs (Taylor, 2012), we first evaluated the importance of these pathways in AGEs clearance in vivo by monitoring levels of MGO-derived hydroimidazolone 1 (MG-H1), one of the most abundant AGEs. Young (3-4 months old) and old (24-26 months old) rats were injected in the hippocampus with the UPS inhibitor lactacystin (Gavilan et al., 2015), and the expression of MG-H1 was analyzed. In the absence of UPS inhibitor, we observed no significant levels of MG-H1-AGEs, hereafter called AGEs, in the hippocampus from young rats and limited AGEs in old rats ( Figure 1a, lanes 1,4). However, there was a significant accumulation of AGEs in old rats when the UPS was inhibited in vivo ( Figure 1a, lanes 5,6 versus 2,3). These data indicate that 1) the combined UPS and autophagic capacity are largely operational at both ages, albeit, 2) AGEs-degrading UPS activity is limited in the older rats. 3) That there is no accumulation of AGEs in UPS-inhibited tissue in young animals is consistent with autophagy compensating for the pharmacological reduction of UPS activity in young but not in old rats (Figure 1a, lanes 5, 6 versus lanes 2, 3) (Gavilan et al., 2015).
We previously found that lysosomal activity was involved in clearance of AGEs, some of which were ubiquitinated, but the mechanism of this process remained an enigma (Uchiki et al., 2012). p62 is an autophagic receptor that recruits ubiquitinated substrates to autophagosomes for subsequent degradation in the autolysosome (Pankiv et al., 2007). This suggested the hypothesis that p62 plays a role in AGEs clearance. A role for p62-dependent autophagy in removal of AGEs is further suggested by the observation that the autophagic receptor p62 and its active phosphorylated form Ser403 (Matsumoto et al., 2011) are upregulated in hippocampus in vivo at versus 4), indicate that autophagy is not upregulated in older animals (Gavilan et al., 2015).
A vast literature indicates accumulation mainly in the insoluble fraction of autophagic cargoes when p62-selective autophagy is impaired (Komatsu et al., 2007), but this has not been explored with regard to AGEs. The retina and lens accumulate AGEs with age as their proteolytic capacities decline (Uchiki et al., 2012). We observed increased levels of AGEs in cells derived from these organs upon prolonged lysosomal blockage by exposure to chloroquine (CQ), which inhibits autophagosome-lysosome fusion and lysosomal acidification without negatively affecting UPS function (Figure 1b, Figure   S1a,b) (Wang et al., 2013). High levels of endogenous AGEs were ob- and Figure S1 c-e). Importantly, in both retinal pigment epithelial (RPE) and human lens epithelial cells (HLEC), we also found that AGEs colocalized with p62 in autophagic vesicles upon CQ treatment (Figure 1f and Figure S2 a-c).
Overall, our data suggest a potential role for p62 in the targeting of endogenous AGEs to the autophagosomal compartment for degradation. Our study also shows that a deficit in the autophagic/ lysosomal function results in accumulation of insoluble AGEs that could compromise cell viability.

| Loss of the autophagic receptor p62 leads to glycation-derived toxicity and AGEs accumulation in vitro and in vivo
To assess the functional ramifications of p62-dependent accumulation of AGEs, we tested if p62 plays a protective role against glycation-derived toxicity. MGO, the primary biologic glycating metabolite of glucose reagent, leads to glycative stress and in vitro accumulation of AGEs (Uchiki et al., 2012). Mouse embryonic fibroblasts (MEFs) derived from wild-type (WT) and p62 knockout animals were exposed to increasing concentrations of MGO for 24 hours (longterm treatment). Concentrations of MGO ≤0.5 mM only caused a 10% decrease in cell viability in p62+/+ cells, while p62−/− cells displayed a fivefold greater susceptibility to glycation-induced toxicity    Figure S3). In the whole body p62 knockout mouse, there was a 120% higher level of MG-H1 in liver ( Figure S3a,b). In the liver-specific p62 knockout mouse, there was a 15% increase MG-H1 ( Figure 2c). The RPE is a major site of AGEs accumulation that is linked to AMD pathology (1, 27, 28). There was a 17% increase in AGEs in the RPE of 12-monthold whole body p62−/− mice ( Figure 2d). We even observed a trend for accumulation of AGEs in the RPE of 3-month-old p62 knockout mice compared with age-matched WT controls that were fed normal diets ( Figure S3c,d). Further generalizing these observations, we also found 58% higher levels of AGEs in whole body Caenorhabditis elegans lacking the p62 ortholog, sqst-1, compared with WT animals Together, these findings suggest that once glycation of a protein reaches a threshold, degradation of AGEs is not efficient, insolubilization or aggregation ensues, and this is associated with accelerated cytotoxicity under glycative stress.

| p62-dependent lysosomal targeting is compromised upon glycative stress
Our observation of cytotoxicity and accumulation of AGEs upon glycative stress suggested that p62 and its function might also be victims of such stress, which, in turn, would then limit the targeting of AGEs to the autophagosome. Since p62 is trapped along with cargo in autophagic vesicles and degraded, accumulation of p62positive vesicles upon lysosomal blockage is an indicator of autophagic flux. Removal of serum from cells upregulates autophagy and was used to explore the fate of p62 in two different cell types highly sensitive to glycative stress and highly active for autophagy (Bejarano et al., 2012;Uchiki et al., 2012). As expected, inhibition of lysosomal function by CQ resulted in an increase in the num- How might glycation interfere with transfer of p62 to autophagosomes? Phosphorylation of p62 at Serine 403 (S403) enhances its ability to recognize substrates and associate with autophagosomes (Matsumoto et al., 2011). We observed that glycative stress rapidly reduced the phosphorylated form of p62; phosphorylation levels fell to 50% after 2 hours of MGO treatment. (Figure 4h,i). Together, these findings indicate that targeting of p62 to lysosomes is compromised upon glycative stress.
In order to determine if p62 suffers the same fate as other proteins upon glycative stress, including crosslinking and aggregation, we specifically monitored p62 fate. Upon short incubation of MEFs with MGO, p62 accumulated in high molecular weight forms F I G U R E 1 Suppression of lysosomal degradation leads to accumulation of endogenous AGEs in autophagosomes. (a) Representative Western blots for MG-H1 in young (3-4 months old) and aged (24-26 months old) rat hippocampus after proteasome inhibition. Note the increased amount of MG-H1 in the aged group at 6 h and 14 h after lactacystin-injection. p62 and phospho-p62 are shown as autophagic markers and GAPDH as loading control. (b) ARPE-19 maintained in the presence or absence of CQ for either 24 or 48 h were subjected to extraction with 1% Triton X-100. Soluble (left) and insoluble (right) fractions were immunoblotted for the indicated proteins. (C,D) Quantification of total soluble (c) and insoluble (d) MG-H1 relative to values in untreated cells. Values are mean ±SEM (n = 4). *p < 0.05 and ***p < 0.001 in one-way ANOVA followed by Dunnett's multiple comparison test. (e,f) Accumulation of MG-H1 in autophagosomes. HLECs were maintained in the presence or absence of CQ for 24 h, fixed in cold methanol. (e) anti-LC3 (green) or (f) anti-p62 (green) was used to stain autophagosomes along with anti MG-H1 (red) to detect endogenous AGEs. Red and green channels are shown in black and white in the upper panels for a better visualization. Full fields for panel e and f are shown in SI Appendix Figure S1c  These observations also reflect in vivo experience. AGEs accumulate in the RPE layer and are involved in the AMD pathogenesis that is observed in mice that consume higher glycemic index (GI) diets and that model human AMD (Rowan et al., 2017). p62 levels increased 59% in the RPE of aged mice fed a high-GI diet (Figure 5e-f).
Statistically significant differences were not observed in neuroretina between high-GI diet and low GI diet (Figure 5g), suggesting that the RPE is more sensitive than other ocular tissues to glycative stress and corroborating observations that the RPE is the nidus of AMD pathology (Rowan et al., 2017;Weikel et al., 2012). As in the cell culture experiments, switching from the high GI diet to the low GI diet lowered p62 to basal homeostatic levels ( Figure

| Enhanced autophagy protects against glycative damage in vitro and in vivo
Next, we tested if enhancement of autophagy might promote cell survival upon glycative stress. Autophagy was pharmacologically enhanced using rapamycin. This binds to and inhibits mTOR, mimicking nutrient depletion and stimulating autophagy. As with p62+/+ versus p62−/− MEFs and long exposure (24 hours) to MGO (Figure 2), we observed increased viability in RPE cells exposed to rapamycin compared to control cells at MGO concentrations greater than 1 mM The results were replicated in HLECs (Figure 6c,d).
Of note, this protective effect of autophagy enhancement was observed only in cells replete with p62. That is, we observed that rapamycin was protective in control cells, but not in MEFs lacking p62 ( Figure 6e).
The protection by p62-dependent autophagy against glycative stress was then tested in vivo. Overexpressing the p62 ortholog SQST-1 in C. elegans, which induces autophagy (Kumsta et al., 2019), reduced the levels of endogenous AGEs 70% (Figure 6f). Of note, accumulation of AGEs was greatest in C. elegans from which p62 was deleted and least in C. elegans in which p62 was overexpressed ( Figure S2e). These findings further support a conserved and critical role of this autophagic receptor in the maintenance of non-toxic homeostatic AGEs levels. F I G U R E 2 Lack of p62 leads to accumulation of AGEs in vitro and in vivo. (a) Viability of p62+/+ and p62−/− MEFs treated with the indicated concentrations of MGO for 24 h was measured by Cell-Titer assay. Values are mean ±SEM (n = 6). We observed an interaction (p < 0.0001) between the MGO concentration and the genotype using two-way ANOVA analysis. The differences between p62+/+ and p62−/− after the Sidak's multiple comparison test were significant for the 0.25, 0.5, and 1 mM doses of MGO (*p < 0.05 and ***p < 0.001).

| DISCUSS ION
(b) Immunoblot for MG-H1 in whole cellular extracts from WT MEFs (p62+/+) and MEFs lacking p62 (p62−/−). Representative immunoblot (top) and quantification of total levels of MG-H1 relative to values in treated cells with 1 mM MGO (bottom). Values are mean ±SEM (n = 10). We observed an interaction (p = 0.03) between the MGO concentration and the genotype using two-way ANOVA analysis. The differences between p62+/+ and p62−/− after the Sidak's multiple comparison test were significant for the 2 and 4 mM doses of MGO (*p < 0.05, **p < 0.01). Specifically, we were motivated to explore proteolytic capacities that might be harnessed to limit AGEs accumulation because: 1) research to chemically inhibit the formation of these glycotoxins has not yielded clinically useful drugs in human trials (Nenna et al., 2015), 2) AGEs accumulation and a decline in autophagic activity are These effects are exacerbated upon aging. We note here more high mass AGEs in the insoluble fraction at a given concentration of MGO, suggesting that once glycation of a protein ensues, it continues to react locally until insoluble moieties are formed. This is corroborated by the transition to accumulation of higher mass AGEs in all fractions with increasing time (Figure 3). Accumulation of AGEs was also obvious when either the UPS or autophagy was inhibited ( Figure 1) (Uchiki et al., 2012). Since AGEs accumulation was reversed in diverse cells when autophagy was stimulated, and this conferred viability (Figure 6), we hypothesized that specific mechanisms of autophagy are salutary, at least in part, by providing a means to limit AGEs accumulation (Takahashi et al., 2017;Uchiki et al., 2012).
We focused on p62, a major carrier of autophagic cargoes to lysosomes, and itself an autophagic target, because initial experiments when animals were returned to lower glycemic diets after having consumed high glycemic diets for over 6 months. This also shows that the stress, which reduced p62 function, can be reversed. This opens up promising modalities to treat diseases such as AMD by easily achievable, cost-effective, and dietary intervention (Rowan et al., 2017;Taylor, 2012).

Our identification of p62 as a biologically conserved mediator of
AGEs clearance associated with survival and disease suggests that it would be profitable to explore drug targets to enhance its activity or stability (Pankiv et al., 2007). Since there are other mammalian autophagic receptors, such as NBR1, NDP52, TAX1BP, and OPTN, future studies will interrogate their function in removal of AGEs and as targets for pharmacotherapy (Birgisdottir et al., 2013).
In sum, we propose a model in which basal autophagic activity contributes to the clearance of endogenous AGEs that are formed through the metabolism of sugars and, additionally, that F I G U R E 3 Absence of p62 leads to higher sensitivity against glycation-derived burden. (a-c) WT MEFs (p62+/+) and MEFs lacking p62 (p62−/−) were incubated with the indicated concentration of MGO for 2 hours, and lysates were separated into 1% Triton X-100 soluble and insoluble fractions. (a) Soluble and insoluble fractions were immunoblotted for AGEs. Quantification of (b) soluble and (c) insoluble MG-H1 in p62−/− MEF relative to values in 1 mM MGO treated p62+/+ cells. Values are mean ±SEM (n = 7). We observed an interaction (p < 0.01) between the MGO concentration and the genotype using two-way ANOVA analysis only for the insoluble fraction (c). The differences between p62+/+ and p62−/− after the Sidak's multiple comparison test were significant for the 2 mM doses of MGO in the insoluble and 1 mM in the soluble fraction (**p < 0.01 and ***p < 0.001). (d-f) Same cells were incubated with 1 mM of MGO for indicated times. (d) Representative immunoblot and quantification of (e) soluble and (f) insoluble MG-H1 relative to values in 2 h treated p62+/+ cells. Values are mean ±SEM (n = 8). We observed an interaction (p < 0.01) between the time of MGO treatment and the genotype using two-way ANOVA analysis for the soluble and insoluble fraction. *p < 0.05, **p < 0.01 and ***p < 0.001. The differences between p62+/+ and p62−/− after the Sidak's multiple comparison test were significant for the 2, 4, and 6 h of MGO in the soluble fraction and for 4 and 6 h of MGO in the insoluble fraction p62-dependent autophagy participates in safeguarding cells and tissues in response to AGEs overload ( Figure S5). Furthermore, at levels of glycative stress that leave p62 functional, it is a major executor of clearance of AGEs. At elevated levels of glycative damage, p62 is rendered dysfunctional. Collectively, our findings suggest that The specific sensitivity of the RPE as a site of accumulation of AGEs is consistent with the RPE being the nidus of AMD. Since glycation has been shown to affect and enhance aggregation patterns of cataractogenic lens as well as neurotoxic α-synuclein and tau proteins associated with Parkinson's and Alzheimer's disease, respectively, these findings suggest that p62-dependent autophagy induction may be salutary with regard to these widely prevalent pathologies as well (Emendato et al., 2018;Liu et al., 2016;Vicente Miranda et al., 2016).

| Animal husbandry
C57BL/6 J p62 knockout mice (allele designation is Sqstm1 tm1Keta ) and liver-specific p62 knockout were donated by Dr. Komatsu Masaaki (Juntendo University, Japan), and WT mice were fed regular chow diet for either three or twelve months. Details regarding the high-GI (HG) or low GI (LG) diet-fed mice can be found in (Rowan et al., 2017). In brief, C57BL/6 J WT mice were purchased from Jackson Laboratories and fed standard chow until 12 months of age. Then, the mice were placed on either HG or LG diets and were pair fed.
At 18 months of age (6 months after the diets were commenced), half of the HG mice were changed over to the LG diet (HGxoLG).  was inhibited by addition of 30 μM CQ, enhanced autophagy was achieved with 1 µM Rapamycin, and glycative stress was induced by addition of MGO. All the studies with MGO were performed at cell density of 90% confluency, as previously described (Uchiki et al., 2012). The fluorometric CellTiter-Blue Assay from Promega was used for assessing cell viability.

| Antibodies and chemicals
The following primary antibodies were used in the study: p62 LG HG HGxoLG LG HG HGxoLG was produced in this laboratory (Shang & Taylor, 1995). Texas Red-and FITC-conjugated secondary antibodies were purchased from Jackson ImmunoResearch Laboratories. Secondary antibodies against mouse and rabbit conjugated to HRP were purchased from Vector Laboratories. Methylglyoxal (MGO) and chloroquine (CQ) were obtained from Sigma-Aldrich. Penicillin-streptomycin solution and FBS, non-essential amino acids, and sodium pyruvate were from GIBCO and protease inhibitor cocktail from Roche.
Rapamycin was purchased from LC Laboratories, USA.

| Detergent solubility assay and immunoblot analysis
Cells were rinsed with phosphate-buffered saline (PBS) at 4°C, collected and centrifuged at 1,000 × g for 5 min. Total whole cellular extracts were prepared by resuspending the cellular pellets in PBS with protease inhibitors, sonicated, and the amount of protein in the samples was estimated using BCA Protein Assay Kit (Pierce). The detergent solubility assay with 1% Triton X-100 was performed as described previously (Bejarano et al., 2012

| Immunostaining and image analysis
Indirect immunofluorescence was performed following conventional procedures, as previously described (Bejarano et al., 2012;Rowan et al., 2017). Briefly, cells were grown on coverslips, fixed for 10 min in either ice-cold methanol or 4% formaldehyde in PBS, blocked and permeabilized 10 min with PBS containing 0.5% BSA, 0.01% Triton X-100 and then incubated with the primary antibody followed by corresponding Alexa 488-or Texas Red-conjugated secondary antibodies. After immunostaining, cells were rinsed with PBS and mounted for microscopy using Fluoromount-G containing DAPI to highlight the nuclei. Immunohistochemistry was carried out as previously described (Bejarano et al., 2012;Rowan et al., 2017;Bejarano et al., 2018). activity, a quantitative analysis was performed according to the guidelines for the use and interpretation of assays for monitoring autophagy (Klionsky et al., 2016). The number of fluorescent puncta and area occupied by p62-positive puncta per cell was calculated using ImageJ.
Subtraction between densitometric values in presence of CQ and in F I G U R E 5 Accumulation of high molecular weight p62 upon glycative stress is reversible. (a,b) WT MEFs (p62+/+) and MEFs lacking p62 (p62−/−) were incubated with the indicated concentration of MGO for 2 hours, and whole cellular extracts were immunoblotted against p62. (a) Representative immunoblot and (b) quantification of p62 monomer and high molecular weight p62 (HMW-p62) values relative to untreated cells are shown. Values are mean ±SEM (n = 5). We observed an interaction (p < 0.0001) between the MGO concentration and the HMW-p62 using two-way ANOVA analysis. The differences between HMW-p62 and monomeric p62 after the Sidak's multiple comparison test were significant for the 4 mM doses of MGO (***p < 0.001). (c,d) ARPE-19 cells were treated with 2 mM MGO for 2 hours followed by incubation in complete medium (no MGO) for either 2 or 4 hours. Cellular lysates were subjected to extraction with 1% Triton X-100 and soluble and insoluble fractions were immunoblotted for the indicated proteins. (c) Representative immunoblot and (d) quantification of HMW-p62 values relative to untreated cells are shown. Values are mean ±SEM (n = 5). Differences between t0 and insoluble p62 were significant for the 4 mM doses of MGO using one-way ANOVA followed by Dunnett's multiple comparison test (**p < 0.01). (e,f) Retinal tissue sections from low-glycemic (LG), high glycemic (HG), and crossover diet (HGxoLG) were analyzed immunohistochemically for p62.

| Image and statistical analysis
Densitometric quantification of immunoblots was performed in unsaturated images using ImageJ (NIH). All Western blot data were was used for analysis of statistical significance. One-way ANOVA followed by Dunnett's multiple comparison test was used when values were compared to untreated controls. Two-way ANOVA followed by the Sidak's multiple comparison test was used when we compared selected pairs of means. Three-way ANOVA followed by Tukey's multiple comparison test was carried out to analyze the effect of genotype, MGO dose and rapamycin in Figure 6e. Two-tailed Student's t test was used to evaluate single comparisons between different experimental groups. Differences were considered statistically significant for a value of p < 0.05 and denoted by an asterisk in the graph.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest. the data, coordinated the study, and wrote the paper.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing not applicable to this article as no datasets were generated or analyzed during the current study. F I G U R E 6 Enhanced autophagy protects against glycative damage by reducing AGEs accumulation. (a, b) ARPE-19 cells were treated with the indicated concentrations of MGO in the absence or presence of 1 µM rapamycin. (a) Cell viability was measured by Cell-Titer assay. Values are mean ±SEM (n = 7). We observed significant effects of both the MGO concentration and the rapamycin using two-way ANOVA analysis (p < .00001). The differences after rapamycin treatment were significant for all the doses of MGO after the Sidak's multiple comparison test (*p < 0.05, **p < 0.01). (b) Immunoblot against MG-H1 is shown. (c, d) HLECs cells were treated under the same conditions. (c) Cell viability, and (d) immunoblot against MG-H1 are shown. Values are mean ±SEM (n = 7). We observed interaction between the MGO concentration and rapamycin using two-way ANOVA analysis (p = 0.0035). The protective effect of rapamycin treatment on cell survival was significant for the 2 and 4 mM doses of MGO after the Sidak's multiple comparison test (**p < 0.01 and ***p < 0.001). (e) WT MEFs (p62+/+) and MEFs lacking p62 (p62−/−) were incubated with the indicated concentrations of MGO in the absence or presence of 1 µM rapamycin and cell viability was analyzed. Values are mean ±SEM (n = 7). We analyzed the effects of p62 genotype, MGO dose, and rapamycin using 3-way ANOVA matching by MGO dose and rapamycin. The three factors have significant effect: p62 genotype (*p < 0.05), MGO dose (***p < 0.001), and rapamycin (**p < 0.01). The only significant interaction was between rapamycin and p62 genotype (p = 0.0077), because it has protective effect only in p62+/+ MEFs. (F) Immunoblot against AGEs in WT and p62-overexpressing C. elegans. Representative immunoblot (left) and quantification of total levels of MG-H1 relative to values in WT (right) Values are mean ±SEM (n = 3). *p < 0.05, **p < 0.01 and ***p < 0.001