Age-Related Defects in Erythrocyte 2,3-Diphosphoglycerate Metabolism in Dementia

Abstract

Alzheimer disease (AD) is the most common dementing illness. Metabolic defects in the brain with aging contribute to the pathogenesis of AD. These changes can be found systematically and thus can be used as potential biomarkers. Erythrocytes (RBCs) are passive “reporter cells” that are not well studied in AD. In the present study, we analyzed an array of glycolytic and related enzymes and intermediates in RBCs from patients with AD and non-Alzheimer dementia (NA), age-matched controls (AC) and young adult controls (YC). AD is characterized by higher activities of hexokinase, phosphofructokinase, and bisphosphoglycerate mutase and bisphosphoglycerate phosphatase in RBCs. In our study, we observed that glycolytic and related enzymes displayed significantly lower activities in AC. However, similar or significantly higher activities were observed in AD and NA groups as compared to YC group. 2,3-diphosphoglycerate (2,3-DPG) levels were significantly decreased in AD and NA patients. The pattern of changes between groups in the above indices strongly correlates with each other. Collectively, our data suggested that AD and NA patients are associated with chronic disturbance of 2,3-DPG metabolism in RBCs. These defects may play a pivotal role in physiological processes, which predispose elderly subjects to AD and NA.

Alzheimer disease (AD) is a classic example of a pathological condition in which learning, memory and cognitive functions decline simultaneously, dramatically and permanently [1]. This chronic illness progresses quite slowly, for many years, manifesting a variety of neurological and psychiatric disorders. Currently, AD affects approximately 20–30 million people worldwide, underscoring the need for brain-specific treatments or preventive interventions [2]. So far, there are no definitive premortem diagnostic tools and no effective therapeutics available for AD. None of AD patients have been cured by any means for a century, and thus AD remains undefeated. Therefore, modern doctrines of AD such as cerebral AD localization and amyloid cascade hypothesis, or at least one of these, can be invalid. Such interpretation provides a platform for investigating AD pathology on the new experimental basis. In the present study, we focus our attention on the metabolic changes in “unusual” erythrocytes rather than in the “accustomed” brain as the new basis for research in AD. We aimed to elucidate whether brain-born AD is invariably associated with RBC function. Growing evidence indicates that in AD, not only brain but peripheral tissues and cells are also affected [3], including blood cells (RBCs, platelets, lymphocytes) [4–6].

RBC is the oxygen carrier, so its metabolic status is important for the regulation of oxygen affinity of hemoglobin (Hb). The ability of Hb to release O₂ is determined by a variety of metabolites in RBCs, especially 2,3-diphosphoglycerate (2,3-DPG), a by-product of glycolysis unique to RBCs. 2,3-DPG is synthesized in the Rapoport-Luebering glycolytic shunt, which bypasses the phosphoglycerate kinase (PGK, EC 2.7.2.3) step. Two components of the shunt, the anabolic BPGM (EC 5.4.2.4) and catabolic BPGP (EC 3.1.3.13) reactions, accounts for the regulation of the 2,3-DPG level [7].

RBCs from AD subjects display altered cell form and impaired deformability [8, 9], suggesting disturbance of ion balance across the plasma membrane and altered RBC ability to pass through the microcirculation. An increase in ion pumping is reflected by increased ATP hydrolysis and enhanced rate of glycolysis. Most surprisingly and paradoxically, so far no one has correlated AD with changes in RBC glycolytic energy metabolism.

The aim of this work was to test the hypothesis that “RBC glycolytic metabolism changes in AD”. We, therefore, measured activities of twelve glycolytic, pentose phosphate shuttle, and Na+/K+-ATPase (EC 3.6.3.9) enzymes and 2,3-DPG levels in RBCs from AD and non-Alzheimer’s dementia (NA) patients, age-matched controls (AC) and young adult controls (YC).

MATERIALS AND METHODS

Patients

We studied RBCs from all AD patients and most NA patients examined at the Pushchino Medical Center during 2005–2008. A total of 59 subjects aged 25–88 years were enrolled between 2005 and 2008. Each patient admitted to the hospital was suffering from dementia.

Body mass index stability and blood test biochemical data were the main criteria for inclusion. All subjects lived at home and did not receive antipsychotics or antidepressants for at least two months. Control subjects had no history of chronic diseases, neurological and psychiatric disorders or treatment. Patients and control subjects entered the investigation after obtaining informed consent from either the subjects themselves or their family members. The Medical Social Examination Commission of Serpukhov Neurological Clinic verified the clinical diagnosis via pathological confirmation for all patients using CERAD criteria [10, 11]. Patients that did not meet the criteria for AD were referred to as NA patients. They had vascular, alcoholic, cranial trauma and other types of dementia. We analyzed different NA patients together. All patients were diagnosed with AD or NA for the first time. Those from the AD group showed no changes in apolipoprotein E ε4 allele and could be considered as demented patients with the symptomatic early phase of late-onset AD. All individuals participated in the study only as blood donors.

Fifteen AD patients were enrolled initially but three patients were later excluded due to appreciable loss of body mass. One NA patient with high blood urea and creatinine, and two YC subjects with high glucose levels were also excluded. After these exclusions, 53 subjects remained eligible for the study. We examined twelve AD patients (aged 66–88 years, mean 75), thirteen NA patients (aged 66–82 years, mean 76), fourteen AC persons (aged 66–81 years, mean 77), and fourteen YC persons (aged 25–45 years, mean 33). Some characteristics of the patients and controls are detailed in Table 1.

Table 1.

Some baseline, clinical and biochemical parameters of studied subjects

Clinical parameters	YC	AC	AD	NA
Mean age, years	33.3 ± 3.3	76.8 ± 3	75 ± 2.6	76 ± 2
Sex ratio (women : men)	10 : 4	10 : 4	7 : 5	9 : 4
Body mass index, kg/m2	24.65 ± 1.07	22.39 ± 0.91	21.35 ± 0.91	21.36 ± 0.87
Blood glucose, mM	4.08 ± 0.50	5.14 ± 0.24*	4.82 ± 0.27	5.11 ± 0.22*
Plasma urea, mM	6.3 ± 0.1	6.33 ± 0.46	7.66 ± 0.91	8.09 ± 0.90
Plasma creatinine, μM	74.40 ± 5.49	81.30 ± 3.43	83.71 ± 6.53	82.45 ± 2.88
Plasma total cholesterol, mM	3.50 ± 0.24	4.98 ± 0.60	5.04 ± 0.39	4.87 ± 0.25
Serum β-lipoprotein, units	33.30 ± 1.23	30.90 ± 2.26	24.92 ± 1.26*	24.23 ± 1.01**
RBC count, 1012/1	4.51 ± 0.07	4.44 ± 0.22	4.73 ± 0.21	69 ± 0.24
White-cells count, 109/l	6.73 ± 1.2	6.43 ± 0.9	5.72 ± 0.8	09 ± 1.0
Hemoglobin, g/l	142 ± 3	139 ± 7	147 ± 9	40 ± 12
Hematocrit, %	40.9 ± 0.9	41.1 ± 1.5	42.2 ± 2.0	41 ± 1.2
Blood aspartate aminotransferase, U/l	20 ± 3	22.4 ± 4.1	16.2 ± 3.1	2.0 ± 3.5
Blood alanine aminotransferase, U/l	22.2 ± 3.4	17.9 ± 4.7	18.8 ± 4.7	20.0 ± 4.2

Results are the mean ± standard error. Values that are significantly different from those of YC group are indicated by asterisks.

^*p < 0.05,

^**p < 0.01 as compared with YC (ANOVA with Bonferroni’s multiple comparison test).

YC, young adults control; AC, age-matched control; AD, Alzheimer’s disease; NA, non-Alzheimer’s dementia.

The investigation conformed to the principles outlined in the Declaration of Helsinki. The institutional review boards at the Medical Center and Institute of Theoretical and Experimental Biophysics approved this investigation before its implementation.

Erythrocyte preparation

Venous blood for clinical use was drawn at the time of admission between 8:00 and 9:00 a.m. after an overnight fast. Five hundred microliters of citrate-anticoagulated blood was used for assays in research laboratory. The washed RBCs were treated as recommended by the International Committee for Standardization in Haematology [12]. Procedures for RBC treatment, preparation of hemolysate and the membrane and cytosolic fractions are described elsewhere [13, 14].

Clinical and standard biochemical analyses were performed immediately (within 30 minutes) after blood collection in the hospital at the Pushchino Research Center. Preparation and biochemical analyses of RBCs were performed (beginning within following 10–20 minutes) in the Institute of Theoretical and Experimental Biophysics. The two groups of buildings, the medical center and research institute, lie about 200 meters apart. Erythrocyte suspensions were stored at 4°C for no longer than 24 h.

Preparation of hemoglobin-free erythrocyte membranes

Washed and packed RBCs with a hematocrit of approximately 85% were lysed in 40 volumes of 5 mM Tris-HCl buffer, pH 8, containing 0.5 mM EDTA and 0.2 mM phenylmethylsulfonyl fluoride at 4°C for 10 min. After centrifugation at 20,000xg for 20 min at 4°C, the precipitate was washed two times with lysing solution and 2–3 times with 100 mM histidine-imidazole-EGTA buffer, pH 7.2. The white sediment was resuspended in the same buffer to give protein concentration of 2 mg/ml and immediately used for Na+/K+-ATPase assay.

Preparation of acid extracts

Washed RBCs (0.2 ml) were mixed with 1 ml of the cold (−20°C) 6% HClO₄/40% ethanol mixture (v/v) and centrifuged at 10,000xg for 10 min at −10°C. The supernatant was neutralized with dry KHCO₃ and 30% KOH. Samples were kept in an ice bath (−20°C) under continuous stirring. Immediately after the centrifugation under above conditions, KClO₄ crystals were removed and a clear supernatant was used to determine metabolites.

Enzyme assay

The washed RBCs were treated as recommended by the International Committee for Standardization in Haematology [12] and enzyme assays were performed on the hemolysate. Enzyme activities were analyzed immediately or after storage of the samples for not more than 24 h at −20°C.

We determined the activities of glucose-6-phosphate isomerase (GPI, EC 5.3.1.9), glyceraldehyde-3-phosphate dehydrogenase (GAPDH, EC 1.2.1.12), PGK, BPGM, BPGK, glucose-6-phosphate dehydrogenase (G6PDH, EC 1.1.1.49) and 6-phosphogluconate dehydrogenase (6PGDH, EC 1.1.1.44) in the hemolysate essentially by Beutler’s methods [15]. Briefly, the components contained in the reaction mixtures for the assays were as follows:

GPI. 0.1 M Tris-HCl, pH 8, 0.5 mM EDTA, 10 mM MgCl₂, 0.2 mM NADP, 0.1 U/ml of G6PDH; the reaction was started by adding hemolysate and 2 mM fructose-6-phosphate.

GAPDH. 0.1 M Tris-HCl, pH 8, 0.5 mM EDTA, 10 mM MgCl₂, 0.2 mM NADH, 8 mM ATP, 5 U/ml of PGK; after 10 min pre-incubation with hemolysate, the reaction was started by adding 10 mM 3-phosphoglycerate.

PGK. 0.1 M Tris-HCl, pH 8, 0.5 mM EDTA, 10 mM MgCl₂, 0.2 mM NADH, 8 mM ATP, 4 U/ml of GAPDH; after 10 min pre-incubation with hemolysate, the reaction was started by adding 10 mM 3-phosphoglycerate.

BPGM. 0.1 M Tris-HCl, pH 8, 0.5 mM EDTA, 1 mM NAD, 5 mM fructose-1,6-diphosphate, 1 mM 3-phosphoglycerate, 20 mM KH₂PO₄, 1 U/ml of GAPDH, 0.1 U/ml of fructose-1,6-bisphosphate aldolase (EC 4.1.2.13), and 1.2 U/ml of triosephosphate isomerase (EC 5.3.1.1); the reaction was started by adding hemolysate.

G6PDH. 0.1 M Tris-HCl, pH 8, 0.5 mM EDTA, 10 mM MgCl₂ and 0.2 mM NADP; after 10 min preincubation with hemolysate, the reaction was started by adding 0.6 mM glucose 6-phosphate.

6-PGDH. 0.1 M Tris-HCl, pH 8, 0.5 mM EDTA, 10 mM MgCl₂ and 0.2 mM NADP; after 10 min preincubation with hemolysate, the reaction was started by adding 0.6 mM 6-phosphogluconate.

Assay of BPGP activity was carried out using the method of Calvin et al. [16]. The components contained in the reaction mixtures for the activity assay were 50 mM triethanolamine-HC1 buffer, pH 7.5, 10 mM MgCl₂, 3 mM ATP, 0.2 mM NADH, 0.8 mM 2,3-DPG, 3.3 U/ml of GAPDH, 2.3 U/ml of PGK, and 1 mM 2-phosphoglycolate. Hexokinase (HK, EC 2.7.1.1), phosphofructokinase (PFK, EC 2.7.1.11), pyruvate kinase (PK, EC 2.7.1.40), and lactate dehydrogenase (LDH, EC 1.1.1.27) activities were measured in the preliminary work [17].

Assays were carried out at 37°C and 340 nm in a Kontron (Uvikon 923) double-beam recording spectrophotometer and Varian Cary Eclipce fluorescence spectrophotometer. Enzymatic activities were determined spectrophotometrically by measuring the rate of NADPH or NAD+ formation.

Activity of Na⁺/K⁺-ATPase was determined by a modified method of Norby [18]. Fifty microliters of the RBC membrane fraction was pre-incubated with 0.02% saponin and then added to 1 ml of a buffer solution (30 mM histidine, pH 7.4, 130 mM NaCl, 20 mM KCl, 4 mM MgCl₂, 3 mM ATP, 1 mM phosphoenolpyruvate (PEP), 10 U/ml of PK, 10 U/ml of LDH and 0.18 mM NADH) containing 1 mM ouabain or no ouabain. The enzyme activity was calculated from difference in the rates of the NAD⁺ formation in the absence and in the presence of ouabain and expressed as mU/mg protein of the membrane fraction.

Protein content was determined according to Lowry et al. [19] with bovine serum albumin as standard.

Metabolite determination

2,3-DPG in the RBC acid extract was determined essentially by Beutler’s method [15]. Briefly, the components contained in the reaction mixtures for the assays were 0.1 M Tris-HCl, pH 8, 0.5 mM EDTA, 5 mM KH₂PO₄, 1 mM MgCl₂, 2 mM GSH, 10 mM hydrazine SO₄, 0.1 mM NADH, 2 mM ATP, 2 U/ml of PGK, 6.8 U/ml of GAPDH, 27 U/ml of phosphoglycerate mutase (EC 5.4.2.1), 0.5 mM 2-phosphoglycolate.

Assays were carried out at 37°C and 340 nm spectrophotometrically by measuring NAD(P)H or NAD⁺ formation.

PEP was determined fluorimetrically using the method of Czok and Lamprecht [20].

Statistical analysis

Data were presented as mean values ± standard error (SE). Statistical comparisons were performed by Student’s t-test. One-way ANOVA with Bonferroni post-tests for multiple comparisons was performed using GraphPad Prism version 5.00 for Windows, GraphPad Software, San Diego, CA. Bonferroni-corrected p values are given in the figures and tables.

For correlation analyses, median levels of all the measurements obtained on individuals of the group were calculated. These sectional distributions were compared by the Spearman rank test. The p values of < 0.05 were considered significant.

RESULTS

Characteristics of studied subjects

Table 1 demonstrates some parameters of studied subjects of the four groups that entered the study. Mean age (except for young control) and body mass index differed insignificantly between groups. Serum concentrations of urea, creatinine and total cholesterol were also statistically indistinguishable in all groups. In spite of the fact that glucose level was higher in AC and NA patients than in the YC group, these are not outside the reference intervals.

Among the serum markers, only β-lipoprotein was at significantly lower levels in AD and NA patients than in YC group and 30% below the lower boundary of normative range. The C-reactive protein concentrations varied between groups but did not exceed the upper reference interval (< 8 mg/l, not shown). There was no significant difference between the four groups in the WBC and RBC counts, the hemoglobin, hematocrit, or the blood levels of aspartate aminotransferase and alanine aminotransferase.

Activities of glycolytic, pentose phosphate shuttle, and Na⁺/K⁺-ATPase enzymes

Findings of the study on enzyme activities in RBCs from controls and patients are summarized in Figs. 1 and 2. Fig. 1 shows enzymes with higher activities in AD and NA groups (or those showing the trend to be higher) than in YC group. In Fig. 2, enzymes are shown with activities in AD and NA groups that are not higher than in YC group.

Figure 1.

RBC activity of glycolytic and Na⁺/K⁺-ATPase enzymes in healthy controls and patients. Activity of Na⁺/K⁺-ATPase is expressed as mU/mg protein of the RBC membrane fraction and that of other enzymes as U/ml of packed RBC. Values that are significantly different from controls are indicated by asterisks: *p < 0.05 as compared with YC; ⁺p < 0.05, ⁺⁺p < 0.01, ⁺⁺⁺p < 0.001 as compared with AC (ANOVA with Bonferroni’s multiple comparison test). HK, hexokinase; GPI, glucose-6-phosphate isomerase; BPGM, 2,3-bisphosphoglycerate mutase; BPGP, bisphosphoglycerate phosphatase; PFK, phosphofructokinase

Figure 2.

RBC activity of glycolytic and pentose phosphate shuttle enzymes in healthy controls and patients. Activities of enzymes are expressed as U/ml of packed RBC. Values that are significantly different from controls are indicated by asterisks. **p < 0.01 compared to YC; ⁺p < 0.05 ⁺⁺p < 0.01; ⁺⁺⁺p < 0.001 as compared with AC (ANOVA with Bonferroni’s multiple comparison test). G6PDH, glucose-6-phosphate dehydrogenase; 6PGDH, 6-phosphogluconate dehydrogenase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PGK, phosphoglycerate kinase; PK, pyruvate kinase; LDH, lactate dehydrogenase

The primary outcome was that activities of all enzymes studied was found to be lower and activities of HK, GPI, PFK and GAPDH are significantly lower (p < 0.05) in the AC group than in the YC group.

Activities of HK, GPI, PFK, GAPDH, BPGM, BPGP, PGK, PK, LDH, 6PGDH and Na⁺/K⁺-ATPase were 23.1, 25.3, 27.2, 15.2, 20, 21, 12.9, 14.2, 10, 13.1 and 25.6% lower, respectively, in the AC group as compared to those in the YC group.

HK, GPI, PFK, GAPDH, BPGM, BPGP, LDH and Na⁺/K⁺-ATPase activities was significantly higher in AD and NA groups compared to AC group (by 11.4–86.4%, p < 0.01).

Some enzymes in AD and NA groups exhibited greater activity levels than those in YC group (Fig. 1). Activities of PFK and BPGP was found to be 27% and 30% higher in AD group than in YC group, respectively, while no significant differences were found between these as well as other enzymes in NA and YC groups.

Metabolite content

Fig. 3 and Table 2 show RBC levels of some glycolytic intermediates, the NAD⁺/NADH and ΓPK ratios in both control groups and both patient groups.

Figure 3.

RBC levels of 2,3-diphosphoglycerate in healthy controls and patients. **p < 0.01, ***p < 0.001 as compared with YC (ANOVA with Bonferroni’s multiple comparison test)

Table 2.

Some glycolytic intermediates and their derivatives in RBC from patients and healthy controls

Parameters	YC	AC	AD	ND
PEP, nmol/ml	11.5 ± 0.7	10.7 ± 0.8	10.2 ± 0.4	10.1 ± 0.4
Pyruvatea, nmol/ml	34.6 ± 2.1	24.2 ± 2.4**	44.9 ± 1.9**,+++	45.3 ± 2.0**,+++
Lactatea, nmol/ml	1102 ± 52	937 ± 54*	1294 ± 32+++	1232 ± 36+++
NAD+/NADHa	289 ± 15	232 ± 17	312 ± 11++	333 ± 17+++
ΓPKb	20.3	9.85	18.0	17.3

^a, from [17].

^b, calculated from mean ± SE of terms in the equation. Values that are significantly different from controls are indicated by asterisks:

^*p < 0.05;

^**p < 0.01 as compared with the YC group;

⁺⁺p < 0.01;

⁺⁺⁺p < 0.001 as compared with the AC group (ANOVA with Bonferroni’s multiple comparison test).

The RBC concentration of 2,3-DPG was found to be significantly lower in AC group than in YC group (9.63 ± 0.44 μmol/g Hb vs 11.7 ± 0.41 μmol/g Hb, respectively, p < 0.01) (Fig. 3). It was significantly lower in the AD group by 32% (p<0.001) and in the NA group by 31% (p < 0.001) than in the YC group, and significantly lower in AD and NA groups by 18% (p < 0.05) and 16% (p < 0.05), respectively, than in the AC group. There was no significant difference between AD and NA groups in 2,3-DPG levels.

There was no significant difference between YC, AC, AD and NA groups in RBC levels of PEP (Table 2).

In the preliminary work, we have found that in the AC group, the RBC concentrations of pyruvate and lactate were significantly lower and the NAD⁺/NADH ratio was insignificantly lower than those in the YC group [17]. These parameters were evidently highest in AD and NA groups. They are given in Table 2 in order to calculate the mass-action ratio of the PK reaction, ΓPK. The ratio was a half as many in AC as YC groups, and that in AD and NA groups was at the YC level (Table 2).

RBC levels of organic hydroperoxides were found elsewhere to be 171% (p < 0.01), 244% (p < 0.001) and 263% (p < 0.001) higher in AC, AD and NA groups than in the YC group, respectively (Fig. 4, modified, recalculated and redrawn from data of [21]).

Figure 4.

RBC levels of organic hydroperoxides in healthy controls and patients. **p < 0.01, ***p < 0.001 as compared with YC (ANOVA with Bonferroni’s multiple comparison test). Data from Kosenko et al. [21] study (modified, recalculated and redrawn).

Similar findings were observed earlier [21] with both H₂O₂ and the sum of H₂O₂ plus organic hydroperoxides. RBC levels of H₂O₂ were found to be 33% (insignificant), 58% (p < 0.01) and 56% (p< 0.01) higher in AC, AD and NA groups than in the YC group, respectively (Fig. 5, modified, recalculated and redrawn from data of [21]). The sum of H₂O₂ plus organic hydroperoxides were found to be 94%, 140% and 147% (all p < 0.001) higher in AC, AD and ND groups than in the YC group, respectively (Fig. 6, modified, recalculated and redrawn from data of [21]).

Figure 5.

RBC levels of H₂O₂ in healthy controls and patients. **p < 0.01 as compared with YC (ANOVA with Bonferroni’s multiple comparison test). Data from Kosenko et al. [21] study (modified, recalculated and redrawn).

Figure 6.

The sum of H₂O₂ plus organic hydroperoxides in RBC from healthy controls and patients. ***p < 0.001 as compared with YC (ANOVA with Bonferroni’s multiple comparison test). Data from Kosenko et al. [21] study (modified, recalculated and redrawn).

The trend for the hydroperoxide generation was an increase in aging with no dramatic changes in dementia. There were no significant differences between AC, AD and NA subjects in H₂O₂, organic hydroperoxide and the sum of H₂O₂ plus organic hydroperoxides content of RBC.

Intracellular concentration of GSH in AC, AD and NA groups was found to be significantly lower than that in the YC group [21]. There were no differences between AC, AD and NA groups in GSH levels, as well as in GSSG levels and the GSH/GSSG ratio in RBCs [21].

Proteolytic enzymes

RBC activities of calpain and caspase-3 were preliminary determined in the two patient groups and the two control groups [17]. Fig. 7 (modified, recalculated and redrawn from data of [17]) shows that calpain activity in RBCs from AC, AD, and NA groups was almost three-times than that in the YC group (p < 0.001).

Figure 7.

Calpain activity in RBC from healthy controls and patients. n = 6–10. Significant differences are indicated by p values (One-way ANOVA with Bonferroni’s multiple comparison test). Data from Kaminsky et al. [17] study (modified, recalculated and redrawn).

Similar results were obtained with caspase-3 activity (Fig. 8, modified, recalculated and redrawn from data of [17]). In RBCs from AC, AD, and NA groups, it was also about three-times that in the YC group and the differences were statistically significant (p < 0.001). There were no significant differences between AC, AD and NA subjects in the two enzyme activities.

Figure 8.

Caspase-3 activity in RBC from healthy controls and patients. n = 6–10. Significant differences are indicated by p values (One-way ANOVA with Bonferroni’s multiple comparison test). Data from Kaminsky et al. [17] study (modified, recalculated and redrawn).

Correlations between indices

Correlation analyses using the Spearman rank test showed that RBC concentrations of 2,3-DPG correlated positively and highly significantly with GSH/GSSG ratio, ATP, adenine nucleotide pool size, energy charge, activities of GPx and GLT, and negatively with levels of organic hydroperoxides, H₂O₂, and calpain and caspase-3 activities. Spearman rank correlation coefficients for the median values were, as a rule, rs = 1.0 or rs = −1.0 (p < 0.01). As an example, Fig. 9 demonstrates the strong negative linear correlation between the median 2,3-DPG and hydroperoxide levels (the sum of H₂O₂ plus organic hydroperoxides, Fig. 6) in RBCs from four groups of healthy controls and patients (r = −0.996, p < 0.001).

Figure 9.

Correlation between RBC levels of 2,3-DPG and hydroperoxides (the sum of H₂O₂ plus organic hydroperoxides) in four groups of healthy controls and patients. Each point represents the median of the two terms of a subject group as listed above. Data were calculated and drawn by the Curve Expert 1.38 software.

DISCUSSION

AD is a disease of aging, as age is a clear contributor in 100% of AD cases [22–24]. AD belongs to a class of neurodegenerative disorders with accompanying dementia. Cognitive decline in AD patients distinguishes them from persons with other types of dementia [10]. In the present investigation, we aimed to elucidate the metabolic changes in RBCs during AD, and to see whether “brain disease” AD is associated with the RBC function. We analyzed an array of glycolytic and related enzymes and intermediates in RBC from elderly patients with or without AD and NA.

Glucose metabolism in the brain is known to be disturbed in aging, and severely reduced in late-onset AD [25]. Causes of dysregulation of neuronal glucose metabolism may be multifold, like decrease in HK [26, 27], PFK [28], pyruvate dehydrogenase activities [29], and impairment of the oxygen delivery to neuronal cells [30, 31]. Metabolic defects have been reported to be present before the development of clinical AD symptoms [32].

RBCs are the only oxygen carrier and generate energy to perform their own function almost exclusively through the anaerobic glycolysis. Only a few metabolic changes in RBCs in aging are reported in the literature. HK activity and glycolysis rates in RBCs from elderly persons have been found to be lower than those in young volunteers [33, 34]. There is no study dealing with RBC glycolytic energy metabolism in AD. AD patients have been reported to have higher RBC Na⁺/K⁺-ATPase activity when compared with AC group [35]. Our findings are consistent with these data, and also deal with broader spectrum of biochemical parameters in the same individuals and confirm that pathological alterations in AD patients’ RBCs touches upon universal metabolic processes such as glycolysis, pentose phosphate shunt and ion transport. Activities of all glycolytic, pentose phosphate shunt and Na⁺/K⁺-ATPase enzymes are evidently lower in RBCs from elderly subjects than in the YC group, and evidently higher (apart from G6PDH) in RBCs from AD and NA patients than those in the YC group.

Among 12 enzymes studied, the key glycolytic enzymes, HK and PFK, and the 2,3-DPG cycle enzymes, BPGM and BPGP, as well as Na⁺/K⁺-ATPase in RBC from AD and NA groups displayed maximum relative activities as compared with those from the YC group. It suggests that RBCs “energy” enzymes are adapted to alterations of the cell form and ion balance across the plasma membrane, and the 2,3-DPG cycle does result into an increase in Hb affinity to oxygen, in favor of tissue hypoxia. Thus, the trend for all enzymes of intermediary metabolism including BPGM, BPGP, PFK and HK is decrease in activity in aging and significant increase in activity in dementia.

The redox state of the NAD⁺/NADH system is more reduced in old age as compared with youth, indicating the possible limitation of glycolysis activity (at the GAPDH step) by NAD⁺ levels. In AD and NA, the redox state of the NAD⁺/NADH system is more oxidized than in both the control groups, indicating the absence of glycolysis limitation by NAD⁺ deficiency.

There are lower 2,3-DPG levels in RBCs from the AC group than those in the YC group. This finding is consistent with data obtained a quarter of a century ago [33]. Much lower 2,3-DPG levels were found in both AD and NA groups as compared with the AC group. It is in accordance with reduced oxygen transport efficiency of blood observed in AD patients [8]. The trend for the key oxygenation metabolite, 2,3-DPG, is some decrease in aging and a further decrease in dementia.

No dramatic change of 2,3-DPG metabolic removal at lower glycolytic steps is observed in AD and NA groups, as revealed by insignificant differences between AD and NA groups in the ΓPK value, and PGK, PK and LDH activities. The ΓPK ratio is more favorable for continual running of the PK reaction to pyruvate formation in vivo in RBCs of YC, AD and ND groups, than in cells of the AC group. ΓPK variation reflects quantitatively and correlates with changes in maximum catalytic activity of PK in different groups of subjects. ΓPK is known to be a thermodynamic index of the flux through PK in the steady state while PGK, PK and LDH activities correspond to the maximum velocities of the three reactions and to the amount of enzyme protein.

Therefore, the decrease in cellular 2,3-DPG levels could be secondary to the dysfunction of the 2,3-DPG shunt itself, namely changes in the two enzyme activities, BPGM and BPGP. Indeed, a larger increase was observed in BDGP activity in AD (by 64.5%, p < 0.0001, paired comparison) and NA patients (47.8%, p = 0.0003, paired comparison,) as compared with those in the AC group, than the increase in BPGM activity in these patients (by 56.4%, p < 0.0001, and 36.5%, p < 0.05, respectively, paired comparison).

A decrease in RBC concentration of 2,3-DPG is known to be the result and a predictor of disturbance of the RBC glycolytic pathway, defects in the distal glycolytic enzymes, and lack of sufficient energy [36]. Secondly, it results in the increase in Hb affinity to O₂ [36], a decrease in the ability of Hb to unload O₂, so that O₂ is transferred to the brain less efficiently. We believe that this inefficient transfer is not limited to the brain only but to other tissues as well. Third, a decrease in 2,3-DPG levels lead to a reduction of the tissue O₂ concentration. Chronically, e.g. during human aging from 40 to 70–80 years, above events may occur very slowly but permanently in all cells including those having mitochondria (i.e. all cells other than RBC). Mitochondria are not only primarily ATP generators but also are primary consumers of oxygen. Accumulating evidence suggests that mitochondria are the main source of reactive oxygen species (ROS) in the cell. Mitochondrial failure plays a key role in the generation of ROS [37], resulting in oxidative damage to cellular compartments [38]. When exposed to hypoxia, several cell types have been shown to increase production of ROS derived from the mitochondrial electron transport chain [39], such as dopaminergic cell lines PC12 and CATH.a [40], endothelial cells [41], cardiomyocytes [42], pulmonary arterial myocytes [36], etc. Mitochondria, as the primary generators of ROS in the cell, are ultimately responsible for the pathogenesis of much, if not all diseases. Oxidative stress has come to the forefront of AD as a causal theory [43]. Chang et al. [8] studied the underlying mechanisms responsible for hemorheological abnormalities in elderly patients with AD and concluded that many factors including blood viscosity, plasma viscosity, blood visco-elasticity, RBC deformability, and blood flow resistance may impair the oxygen transport efficiency of blood in AD patients.

Erythrocytes themselves can serve as generators and sources of ROS for other tissues. It has been found that in hypoxic RBCs, superoxide and H₂O₂ production increased and the RBC-derived H₂O₂ was released to the extracellular space [44, 45]. In hypoxia, H₂O₂ is transferred from RBCs to the endothelium in lung capillaries [45].

Regarding oxidative stress parameters of RBCs, there are not too many reports of their analyses in the elderly and results obtained are multi-valued [35, 46, 47]. Significant heterogeneity has been observed between works dealing with RBC activities of superoxide dismutase, gluthathione peroxidase, and catalase in AD and NA [48, 49]. We recently carried out in-depth biochemical analyses on both enzymatic and low-molecular weight pro-oxidants and antioxidants in RBCs from demented patients. We observed changes in activities of catalase, superoxide dismutase, glutathione reductase and G6PDH not related to, rather expounded as age-related abnormalities [21]. In addition, preliminary [21] and present studies showed age-related rise of H₂O₂ and organic hydroperoxide levels in RBCs from all elderly groups with and without dementia.

The RBC energy state, as revealed by parameters of the adenylate system, is altered significantly in AD patients. To our knowledge, only one study was performed on adenylate levels in RBCs from elderly subjects [34], which reported decreased ATP concentrations in RBCs from old persons as compared with young volunteers. Our recent results demonstrated that all elderly groups, e.g. AC, AD and NA, showed significant or tendentious declines in ATP, adenine nucleotide pool size, ATP/ADP ratio, and energy charge when compared with the YC group [21]. Remarkably, no significant difference was observed between AD, NA and AC groups regarding any parameter of the adenylate system. The trend for the key energy parameters, ATP, adenine nucleotide pool size and energy charge, is a decrease in aging without further change in dementia [21].

The trend for the two RBC proteolytic enzymes is a dramatic increase in aging with no further changes in dementia. Increased RBC calpain and caspase-3 activities were observed in 100% of elderly persons, without an overlap, indicating the communion of the age effect.

The strong correlation between a numbers of indices suggested association between corresponding metabolic reactions and processes: depletion of energy stores and 2,3-DPG, hydroperoxide accumulation, inactivation of antioxidant defense system and activation of proteolysis.

Our results may provide the new basis for increased oxidative stress in dementias. A defect in RBC glycolytic energy metabolism results in an impairment of 2,3-DPG synthesis, an increase in Hb affinity to oxygen, and loss of adequate oxygen delivery to tissues. Under some (as yet undetermined) conditions, brain (and other tissue) mitochondrial oxidative phosphorylation is inhibited by oxygen deficiency that results in excessive ROS production. Many years of permanent oxidative stress irreversibly affects all tissues including brain and can lead, or cannot lead, to AD or NA, depending on the extent of 2,3-DPG synthesis deficiency of RBC. Such chain of events is an appendix to list of causes for oxidative stress development that have been enumerated by Obrenovich et al. [50].

The results revealed herein may gather importance in respect to several facts. Firstly RBC metabolic disturbances are clearly depicted in aging, AD and NA. Secondly, most of the RBC glycolytic, pentose phosphate shunt and Na⁺/K⁺-ATPase enzyme activities decrease in aging. The increase in these enzyme activities are characteristics of dementia and are non-specific for AD and NA. Furthermore, the hemoglobin affinity to oxygen increases in aging and further elevates by a greater amount in both demented groups. This parameter isolates the AD and NA groups from the AC group but is virtually equal in both the demented groups. As RBCs from AD and NA patients are similar in majority of enzyme activities and steady state metabolite concentrations, they can be considered to share common metabolic defects in the two degenerative disorders.

Among 37 metabolic parameters studied, only four glycolytic enzyme activities, HK, PFK, BPGM and BPGP, increased specifically and glutathione peroxidase activity decreased specifically in AD. Changes in other enzyme activities and metabolite concentrations in RBCs were not specific to dementia but characteristics for aging. Metabolic features of RBCs can be of diagnostic importance for neurological diseases and biomarkers of AD progression. Thus, the present investigation provides an insight into a link between dementia and RBC function. It also highlights a need to examine the role of RBCs in the pathogenesis of AD.