Glyceraldehyde caused Alzheimer’s disease-like alterations in diagnostic marker levels in SH-SY5Y human neuroblastoma cells

Clinical evidence has implicated diabetes mellitus as one of the risk factors for the development and progression of Alzheimer’s disease (AD). However, the neurotoxic pathway activated due to abnormalities in glucose metabolism has not yet been identified in AD. In order to investigate the relationship between impaired cerebral glucose metabolism and the pathophysiology of AD, SH-SY5Y human neuroblastoma cells were exposed to glyceraldehyde (GA), an inhibitor of glycolysis. GA induced the production of GA-derived advanced glycation end-products (GA-AGEs) and cell apoptosis, glycolytic inhibition, decreases in the medium concentrations of diagnostic markers of AD, such as amyloid β 1-42 (Aβ42), and increases in tau phosphorylation. These results suggest that the production of GA-AGEs and/or inhibition of glycolysis induce AD-like alterations, and this model may be useful for examining the pathophysiology of AD.

Cerebrospinal fluid (CSF) tests have been proposed for the early detection of AD 9 . However, there is currently no evidence to suggest that CSF findings in DM patients are an indicator AD. Aβ 42 and total and phosphorylated tau protein (p-tauT181) in the CSF are useful markers for the diagnosis of AD, in which Aβ 42 concentrations are lower while the tau protein and its phosphorylated form are higher than those in healthy subjects 10,11 . A previous study showed that the intracellular phosphorylation of tau was increased in the cortex of AD patients 12 . The CSF levels of vascular endothelial growth factor (VEGF) and transforming growth factor-β (TGF-β ) were found to be significantly higher in AD patients than in healthy controls 13 . However, limited information is available for the mechanisms underlying CSF and intracellular molecular changes. Although recent findings suggest that cerebral glucose metabolism is impaired from the early stage of AD 14,15 , it currently remains unclear whether reductions in the utilization of glucose lead to the development of AD.
In the present study, SH-SY5Y human neuroblastoma cells were treated with GA, an inhibitor of glycolysis 16,17 , and examinations of cell viability and measurements of the concentrations of Aβ 42, total tau, and p-tauT181 proteins in culture media as well as the phosphorylation ratio of intracellular tau were then performed in order to determine the involvement of glucose metabolism in the pathophysiology of AD.

Results
Toxic effects of GA on SH-SY5Y cells. A microscopic analysis revealed that GA clearly exerted cytotoxic effects on SH-SY5Y human neuroblastoma cells at 0-1 mM (Fig. 1a-d). The WST-8 assay showed that the viability of cells was significantly reduced from 0.7 mM (Fig. 1e). Fluorescent microscopic analyses revealed that cells exhibited green fluorescence with YO-PRO ® -1 dye, which detected early apoptosis in a GA-dose dependent manner ( Fig. 1f-i). On the other hand, the red fluorescence of propidium iodide for necrosis was not observed (data not shown), suggesting that GA-induced cell death was mainly caused by apoptosis.

Characterization of GA-induced cell death in SH-SY5Y cells. A relationship has recently been
reported between the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and apoptotic events 14 . Real-time RT-PCR experiments showed that the expression of GAPDH was significantly increased by the addition of GA (Fig. 2a). These results suggested that GA caused cell toxicity concomitant with increases in the gene expression of GAPDH in SH-SY5Y cells. We then measured GA-AGE levels after the 24-h treatment of SH-SY5Y cells with GA using slot blots. A small band of GA-AGEs was observed in the vehicle control (Fig. 2b). GA dose-dependently increased the production of GA-AGEs in SH-SY5Y cells. The reductions observed in cell viability following a 24 h incubation with GA were significantly restored by the addition of 1 mM (P < 0.05) or 10 mM (P < 0.01) acetoacetate (ACAC), a substrate for the tricarboxylic acid (TCA) cycle (Fig. 2d), whereas lactic acid concentrations were not (Fig. 2c). Glucose consumption, calculated by differences in medium concentrations during the incubation, negatively correlated with lactic acid concentrations (data not shown). These results suggested that GA inhibited glycolysis while ACAC restored cell viability without recovering reductions in glycolysis.
Effects of GA on Aβ42 and the tau protein and its phosphorylated form in SH-SY5Y cells. In order to determine the effects of GA on CSF markers for the development of AD, an aliquot of medium was analyzed by an enzyme-linked immunosorbent assay (ELISA) 24 h after the addition of GA. Aβ 42 in the medium decreased in a GA dose-dependent manner (Fig. 3a). In contrast, GA significantly increased tau and its phosphorylated form, p-tauT181 (Fig. 3b,c) in the medium. The ratio of the phosphorylation of intracellular tau was increased significantly at 0.7 mM (P < 0.05) and 1 mM (P < 0.01) GA (Fig. 3d). In addition, VEGF (Fig. 3e) and TGF-β (Fig. 3f), which are also AD biomarkers, were increased when the concentration of GA added was greater than 0.7 mM.

Discussion
The present study has three salient results: 1) GA induced the formation of GA-AGEs and exhibited cytotoxicity in SH-SY5Y cells.
2) The mechanism underlying GA-induced cell death involved the inhibition of glycolysis and concomitant induction of GAPDH. 3) Changes in the levels of AD biomarkers in GA-treated culture media were consistent with those in the CSF of AD patients.
Epidemiological studies recently reported that the risk of developing AD was higher in DM patients than in the general population 18 . The Rotterdam study, which surveyed more than 6,300 patients, indicated a relationship between DM and AD with a relative risk (RR) of 1.9 19 . Given the recent interest in the relationship between insulin and AD, patients in that study receiving exogenous insulin therapy were at the highest risk (RR 4.3) of developing dementia 20,21 . On the other hand, the Honolulu-Asia Aging Study, which explored a cohort of 2,574 patients, also indicated that the risk of developing AD was 1.8 in DM patients while that of vascular dementia was 2.3 22 . Since AGE levels were previously shown to be upregulated in the brains of diabetic patients with AD 23 , these findings may partly explain the clinical link between DM and AD.
The involvement of AGEs in AD has been suggested in several studies published successively between 1994 and 1995 [24][25][26] . SPs were previously shown to be positively stained by an anti-serum against glucose-derived AGEs (Glu-AGEs) 7 . Münch's group also reported that Glu-AGEs had similar effects to those of Aβ , namely, increases in neurotoxicity and glucose consumption in a SH-SY5Y cell culture system [27][28][29] . In our previous studies 30, 31 , we demonstrated that α -hydroxyaldehydes (glyceraldehyde and glycolaldehyde) and dicarbonyl compounds (glyoxal, methylglyoxal, and 3-deoxyglucosone) contributed to the glycation of proteins. We confirmed that GA-AGEs and glycolaldehyde-derived AGEs (Glycol-AGEs) were strongly neurotoxic in a neuronal culture system 1,8 . The neurotoxicities of these AGE species were stronger than those of Glu-AGEs and N-(carboxymethyl)lysine (CML), two extensively examined AGE species. Moreover, the neurotoxic effects of serum AGEs from diabetic patients on hemodialysis were completely attenuated by the addition of an anti-GA-AGE-specific antibody, but not the antibodies of glycolaldehyde-, methylglyoxal-, glyoxal-, 3-deoxyglucosone-, or glucose-derived AGEs 1,8 . Therefore, GA-AGEs are powerful candidate molecules for neurodegeneration in AD, and, consequently, we focused on the mechanism underlying GA-AGE-induced neurotoxicity in the present study.
On the other hand, a Glu-AGE antibody was shown to react with SPs, mainly with the amyloid core, whereas GA-AGE and Glycol-AGE antibodies showed no immunoreactivity with SPs 32 . These findings suggested that Aβ was glycated by glucose, rather than GA. GA-AGEs are mainly present in the neurons of the hippocampus and parahippocampal gyrus. It was mainly localized in the perikarya of neurons and staining pattern was uniform, differing from the dot-like pattern of Glu-AGE staining 32 . Furthermore, Glu-AGEs were detected in intracellular and extracellular sites, whereas GA-AGEs were only found intracellularly. This discrepancy indicated that the mechanisms underlying the neurotoxicity induced by Glu-AGEs and GA-AGEs differed.
In the present study, we detected GA-AGEs intracellularly in SH-SY5Y cells using slot blots. GA dose-dependently increased GA-AGE levels, as shown in Fig. 2b. We previously investigated which types of AGEs triggered the development of AD pathology using a culture system. We showed that cell viability was markedly decreased by the addition of GA-AGEs 1,8 . In the present study, lactic acid concentrations in culture media, to which 1 mM GA had been added, were approximately 50% lower than those in the control and ACAC recovered cell viability from GA-induced cell death. These results suggested that GA induced the inhibition of glycolysis. In a study using positron emission tomography, cerebral glucose metabolism was found to be lower in AD patients than in healthy controls 14 . Needham et al. reported that GA attenuated glycolysis by inhibiting hexokinase and GAPDH 17 . Furthermore, GAPDH activity was previously reported to be reduced in AD patients 33 . The role of GAPDH in apoptosis has already been described in neurodegenerative disorders 34 . Although GA induced the expression of GAPDH in the present study, we previously reported that GA inactivated GAPDH in a neuronal culture system 8 . ACAC recovered from GA-induced cell death, suggesting that abnormalities in glycolysis may be involved in the neurotoxicity mechanism. GAPDH inactivation due to intracellular GA production further increases intracellular GA concentrations, resulting in increased GA-AGE production and nerve cell toxicity. Thus, GA-AGEs may be one of the general causative agents of the development of neurodegenerative diseases, such as AD. The brain is in direct contact with the CSF. Biological alterations that reflect pathophysiological processes in the brain are reflected in the CSF. Evidence obtained over the past 20 years has indicated that Aβ 42 levels in the CSF of AD patients are significantly lower than those in age-matched healthy elderly controls, whereas total tau and p-tauT181 levels are significantly higher 35 . Furthermore, the levels of other AD biomarkers, such as VEGF 13 and TGF-β 1 36 , were also found to be higher in the CSF of AD patients. Shuvaev et al. reported that glycation was greater in the CSF of AD patients than in age-matched controls 37 . In our study, GA decreased Aβ 42 levels and increased total tau and p-tauT181 levels in culture media and also increased the intracellular levels of total tau, p-tauT181, VEGF, and TGF-β in SH-SY5Y cells. We were unable to determine why GA decreases Aβ 42 levels and increased p-tauT181 levels in culture media. GA may have had effects on the post-translational modifications (e.g., phosphorylation in the early stage) of p-tauT181, but not on the expression and/or secretion of Aβ 42 through glycolysis inhibition. Further studies are needed in order to elucidate the mechanisms underlying GA-induced AD-like alterations in diagnostic marker levels. In addition, Glu-AGEs did not induce cytotoxicity, but caused tau-phosphorylation in a culture system 38 . These findings suggest that high levels of glycation and/or GA may be mimicked by AD CSF alterations and accompanied by numerous neuropathological consequences due to GA-AGEs rather than Glu-AGEs.
GA is derived from two distinct pathways, 1) a glycolytic pathway and 2) fructose metabolism pathway 39 . 1) The glycolytic intermediate glyceraldehyde-3-phosphate is generally catabolized by GAPDH. As discussed above, the addition of GA to neuronal cells led to a decrease in GAPDH activity, suggesting a feed-forward mechanism; the GA-AGE-induced inactivation of GAPDH may further stimulate the generation of GA-AGEs. Further studies are needed in order to elucidate the exact mechanisms underlying GAPDH-related neurotoxicity by GA. 2) Under hyperglycemic conditions, an increase in intracellular glucose has been shown to stimulate the polyol pathway to generate fructose in insulin-independent tissues, including the brain and nerve tissue 40 . Fructose is phosphorylated to fructose-1-phosphate and then catabolized to dihydroxyacetone phosphate and GA by liver aldolase B 41 , and GA promotes the formation of GA-AGEs. Moreover, previous studies reported that aldolase B was not expressed in the rat brain 42 , whereas fructose-1-phosphate cleavage (aldolase) activity has been detected in the human brain 43 .
Although the exact mechanisms underlying the target of GA-AGEs and its downstream signaling pathway currently remain unclear, the measurement of GA-AGE levels in the CSF and/or serum may be a useful biomarker for the early detection of AD. Cell culture, cell counting, cell viability, detection of apoptosis, and culture media analyses.

Materials
Cells were cultured in DMEM on 6-well plates for 24 h, followed by the addition of various concentrations (0, 0.3, 0.7, and 1 mM) of GA, and all assays were performed after a further incubation for 24 h. Cell viability was assessed using Cell counting kit-8 (Dojin, Kumamoto, Japan) according to the manufacturer's instructions. Apoptosis was examined by fluorescent microscopy using Vybrant ® Apoptosis Assay Kit #4 (Invitrogen Corporation, CA, USA), in which early apoptotic cells were detected with green fluorescent YO-PRO ® -1. In medium analyses, an aliquot of medium was centrifuged (600 g, 5 min, 4 °C) followed by recentrifugation of the supernatants (21,000 g, 10 min, 4 °C), and the resulting supernatants were used. Aβ 42, total tau protein, and p-tauT181 protein were examined by ELISA kits (for the tau protein, its phosphorylated form, and Aβ 42, the kit was obtained from Invitrogen Corporation, CA, USA). Lactic acid concentrations were measured by Determiner-LA (Kyowa Medex, Tokyo, Japan). In ACAC experiments, cells were treated with ACAC 15 min before the addition of GA.
Statistical analysis. Data are expressed as the mean ± SD, and were examined by a one-way analysis of variance (n = 3 or n = 6). More than two experiments were performed and similar results were obtained. P values less than 0.05 were considered to be significant.