Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 140–145

Accumulation of plasma 3-deoxyglucosone impaired glucose regulation inChinese seniors: Implication for senile diabetes?

Guorong Jiang a, Lurong Zhang a, Qingling Ji b, Fei Wang a, Heng Xu a, Fei Huang a, Chunxiang Wang a,*a Institute of Traditional Chinese Medicine, Suzhou Traditional Chinese Medicine Hospital, Suzhou, Chinab First Clinical Medical College of Nanjing University of Chinese Medicine, Nanjing, China

A R T I C L E I N F O

Keywords:

3-DG

3-Deoxyglucosone

IGT

Impaired glucose regulation

Senile

Diabetes

A B S T R A C T

Aims: To investigate the impact of increasing accumulation of 3-deoxyglucosone (3-DG) on glucose

regulation in non-diabeteic seniors.

Methods: This research is a 2-year prospective follow-up study. We conducted a HPLC assay to

determine the plasma 3-DG concentrations of 132 non-diabetic retirees of Suzhou. An oral glucose

tolerance test was carried out 2 years after baseline in 16 subjects with continual high plasma 3-DG and

16 control subjects randomly sampled in those with normal plasma 3-DG.

Results: The median plasma 3-DG level of 132 subjects was 43.52 ng/ml (7.89–736.09 ng/ml), of which

47 subjects (36.6%) were beyond 70 ng/ml. A correlation between age and 3-DG was found among people

between 50 and 66 years old (r = 0.408, P < 0.001). The 60–69 years group had a higher 3-DG level than

50–59 years group (P < 0.001). Compared with control group, the continual high plasma 3-DG subjects

had a higher level of FINs (P < 0.05), FBG (P < 0.01), HOMA-IR (P < 0.001), and a lower level of ISI

(P < 0.001) and DI60/DG60 (P < 0.05), as well as a higher incidence of impaired glucose regulation

(x2 = 7.814, P < 0.05).

Conclusions: There was abnormal elevation of plasma 3-DG in non-diabetic seniors, and the increasing

accumulation of plasma 3-DG, which mainly resulted from aging, eventually lead to the impaired

glucose regulation, indicating an association of 3-DG with senile diabetes.

� 2012 Diabetes India. Published by Elsevier Ltd. All rights reserved.

Contents lists available at SciVerse ScienceDirect

Diabetes & Metabolic Syndrome: Clinical Research &Reviews

jo ur n al h o mep ag e: www .e lsev ier . c om / loc ate /d s x

1. Background

3-Deoxyglucosone (3-DG), as an important intermediateproduct of Maillard reaction, is a highly reactive carbonyl product.In addition to exerting its potent ability to form advanced glycationend-products (AGEs), 3-DG has some biological activities itself [1–3]. It has showed that the level of plasma 3-DG was obviouslyelevated in patients with diabetes [4,5] and elevated plasma 3-DGlevel may be one of the risk factors for diabetic microangiopathiccomplications [5,6]. However, it is unknown whether 3-DG isinvolved in the genesis of diabetes. On the other hand, it is deducedthat 3-DG was associated with aging mainly according to thebioactivities and the pathological effects of AGEs, such asimplication for a lot of senile diseases [6–8]. However, thecorrelation between aging and 3-DG has not been revealed. Thestudy on interrelationships among plasma 3-DG, diabetes, aging,and senile diabetes is still in progress.

* Corresponding author. Tel.: +86 512 67872503; fax: +86 512 67872506.

E-mail address: wcx.sz@163.com (C. Wang).

1871-4021/$ – see front matter � 2012 Diabetes India. Published by Elsevier Ltd. All r

http://dx.doi.org/10.1016/j.dsx.2012.09.010

The accumulated evidences have showed that AGEs, andmethylglyoxal (MGO), the other intermediate product of Maillardreaction, could affect glucose metabolic regulation throughinducing insulin resistance which is associated with the genera-tion and progression of diabetes both in vitro and in vivo [9,10].Kiho et al. reported that hexokinase and glucose-6-phosphatedehydrogenase activities were decreased by 3-DG, whichsuggested that 3-DG could inhibit the intake of glucose in theliver [2]. Our previous researches have found that 3-DG couldinduce impaired glucose tolerance and decrease insulin sensitivi-ty in healthy mice [11].

Accordingly, we hypothesize that aging is correlated withplasma 3-DG level, and the increasing accumulation of 3-DG maylead to impaired glucose regulation in non-diabetic elderly, whichis associated with senile diabetes.

In this study, we investigated that: (i) the distribution of plasma3-DG levels of non-diabetic retirees of Suzhou; (ii) the associationsof clinical index (i.e. body mass index (BMI), fasting blood glucose(FBG), and postprandial blood glucose (PBG)) with plasma 3-DGlevels; (iii) the possible effects and mechanism of the increasingaccumulation of plasma 3-DG on glucose regulation with a twoyears prospective study.

ights reserved.

Table 1The distribution of plasma 3-DG of retired non-diabetic Chinese people.

3-DG (ng/ml) Absolute

frequency

Frequency Cumulative

relative frequency

0– 85 64.4 64.4

70– 28 21.2 85.6

140– 6 4.5 90.2

210– 1 0.8 90.9

280– 4 3.0 93.9

350– 1 0.8 94.7

420– 3 2.3 97.0

490– 1 0.8 97.7

560– 1 .8 98.5

630– 0 0 98.5

700– 2 1.5 100.0

Total 132 100.0 100.0

G. Jiang et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 140–145 141

2. Methods

2.1. Subjects

We studied the retirees beyond the age of 50 of women and 55of men (defined as seniors according to the Chinese traditionalrule), who visited the Department of Health of Suzhou TraditionalChinese Medicine Hospital from June to November 2009. Thestudy was approved by the Ethical Committee of SuzhouTraditional Chinese Medicine Hospital, and this study wasconducted according to the principles of the Declaration ofHelsinki.

2.2. Clinical examination

A physical examination including height, weight, andblood pressure was carried out. Height and weight were takenwithout shoes or cap and in light clothing, then BMI wascalculated for each subject using the formula: BMI = weight (inkg)/height (in m) squared. We regarded BMI �24 as overweight,and �28 as obese, according to the criterion established byCooperative Meta-analysis Group of China Obesity Task Force[12]. Early morning initial urine samples and fasting bloodsamples were collected and send to the clinical laboratory of thehospital immediately, and FBG, TG, TC, HDL-C, AST, ALT, UA andScr were measured with Automatic biochemical Analyzer(HITACHI 7600). The PBG was measured 2 h after breakfast.The plasma for 3-DG was collected in specific blood collectiontubes to be immediately precipitated the protein fraction andstored at �80 8C until assayed. The concentration of 3-DG wasmeasured respectively at baseline, 1 year and 2 years afterbaseline with HPLC-FLD by our Institute of Chinese Medicine. Anoral glucose tolerance test (OGTT) was carried out to evaluatethe glucose regulation of the further sampled investigatedsubjects.

2.3. OGTT

Participants underwent an OGTT (1.75 g/kg, maximum 75 g).Blood samples were obtained at 0, 60 and 120 min for themeasurement of glucose and insulin. The b-cell function wasassessed by HBCI and DI60/DG60, while the insulin resistance byHOMA-IR and ISI. All of the indexes were calculated using thefollowing formulas: HBCI = 20 � FINs/(FBG � 3.5) [13]; DI60/DG60 = (I60 � I0)/(G60 � G0) [14]; HOMA-IR = FINs � FBG/22.5(FINs: fasting insulin, FBG: fasting blood glucose) [15]; ISI = 1/(FINs � FBG) [16]. The classification of diabetes mellitus wasbased on the reports brought into order by the second WorldHealth Organization Expert Committee on Diabetes Mellitus in1999 [17]: the diagnostic fasting plasma (blood) glucose valuewas �7.0mmol/L, while postprandial glucose level was�11.1mmol/L; Impaired Fasting blood glucose (IFG) encom-passed values which are above normal but below the diagnosticcut-off for diabetes (plasma � 6.1 to <7.0 mmol/L); impairedglucose tolerance (IGT) was above 7.8 mmol/L, but below11.1 mmol/L.

2.4. Quantification of 3-DG

For determination of plasma 3DG levels, we used high-performance liquid chromatography according to Hasuike et al.[18]. 3-DG was allowed to react with 2,3-diaminonaphthalenethen detectable by fluorescence (excitation, 271 nm; emission,503 nm). The mobile phase consisted of a mixture of a solventsystem (50 mM phosphate, acetonitrile, and methanol = 7:2.5:0.5)at a flow rate of 1.0 ml/min on a reverse-phase.

2.5. Statistical analysis

Data from the subjects were double-checked and cleanedbefore it was entered into a computer. The data for subjects whoseactual measurement values (or log-converted values) showednormal distributions were recorded as the mean � SD, and the datadid not show a normal distribution were recorded as the median(range). We used Student’s t-test for comparing means between twogroups, and one-way ANOVA between groups, followed by S-N-Kmultiple comparison test. The Mann–Whitney and Kruskal–Wallistest were used to compare the medians for the continuous variables,while the chi-squared test was used for the categorical variables. Thecorrelation between 3-DG and the various parameters was studiedusing single regression analysis. The P value of <0.05 was regarded assignificant.

3. Results

3.1. The distribution of plasma 3-DG level in non-diabetic seniors

Some known pathological factors, including a history ofhyperglycosemia or nephrosis, the FBG � 6.1 mmol/L orPBG � 7.8 mmol/L, and abnormal of liver or renal function, mightaffect the plasma 3-DG level. So the subjects with those factorswere excluded in the test.

The number of eligible subjects were 132, including 44 men and88 women, with the average age of 60.73 � 6.797 years (50–83years). It showed a skewed distribution of plasma 3-DG, with themedian of 43.52 ng/ml (7.89–736.09 ng/ml), the P95% of 452 ng/ml,and the 95%CI from 70.40 to 116.25 ng/ml. We graded the level of 3-DG by 70 ng/ml [19], and found that 64.4% �70 ng/ml, while 25% were70–300 ng/ml, and 10.61% �300 ng/ml (Table 1).

3.2. The relationships between the clinical parameters and

plasma 3-DG

We converted the data of 3-DG with napierithm to make anormal distribution. The results showed there was no correlationbetween FBG or PBG and 3-DG (r = 0.129 and 0.134, P = 0.139 and0.123). No difference in plasma 3-DG (3.99 � 0.93 vs. 3.95 � 1.01,P = 0.815) or other parameters was found between males andfemales.

According to their ages, we divided the subjects into threegroups, 50–59 years, 60–69 years, and �70 years group. Asignificant difference of 3-DG level was found among these threegroups (ln(3-DG): 3.53 � 0.79, n = 49 vs. 4.27 � 1.01, n = 69 vs.

3.98 � 0.92, n = 14, F = 9.277, P < 0.001). The 60–69 years group had ahigher level of 3-DG than 50–59 years group (P < 0.001), while peoplebeyond 70 years old had a slightly lower levels of 3-DG, compared

Fig. 1. The relationships between the clinical parameters and plasma 3-DG. (A) The compare of plasma 3-DG level in different age. The 95%CI of 3-DG of three groups (50–59

years, 60–69 years, and �70 years group) were 35.15–57.94 ng/ml, 87.78–168.66 ng/ml, and 27.80–142.42 ng/ml respectively, a significant difference of 3-DG was found

among these three groups (ln(3-DG): 3.53 � 0.79, n = 49 vs. 4.27 � 1.01, n = 69 vs. 3.98 � 0.92, n = 14, F = 9.277, P < 0.001), and a significant difference was found between 50–59

years group and 60–69 years group (P < 0.001)). (B) The correlation between age and plasma 3-DG. A statistically significant correlation between age and 3-DG level was found

among the 50–66 years old people (r = 0.408, P < 0.001). (C) The correlation between BMI and 3-DG, and FBG. The ln(3-DG) of these three groups (BMI < 24 (n = 72), BMI from 24 to

26 (n = 39), and BMI � 26 (n = 21)) were 3.84 � 0.89, 3.87 � 0.89, and 4.52 � 1.08, F = 4.323, P = 0.016, the FBG were 5.04 � 0.47, 5.01 � 0.55, and 5.38 � 0.58, F = 4.060, P = 0.020,

and the PBG were 6.54 � 0.42, 6.53 � 0.51, and 5.73 � 0.48, F = 1.625, P = 0.201. *Compared with the <24 group, P < 0.05; #compared with the 24–26 group, P < 0.05.

G. Jiang et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 140–145142

with people in 60–69 years group (P > 0.05) (Fig. 1A). Further more, astatistically significant correlation between age and 3-DG level wasfound among people between 50 and 66 years old (r = 0.408,P < 0.001) (Fig. 1B). No correlation was found between age andFBG or PBG (r = 0.030 and 0.735, P = 0.062 and 0.481).

According to their BMI, we divided the subjects into threegroups, BMI < 24 (n = 72), BMI between 24 and 26 (n = 39), andBMI � 26 (n = 21). The levels of ln(3-DG) of three groups were3.84 � 0.89, 3.87 � 0.89, and 4.52 � 1.08 respectively. A significantdifference of 3-DG was found among three groups (F = 4.323,P = 0.016). Subjects with BMI � 26 had a higher rate of 3-DG(�70 ng/ml) than the subjects with BMI < 26 (11/21: 52.38% vs.

32/111: 28.83%, x2 = 8.521, P = 0.004). Compared with the other twogroups (BMI < 24 and BMI between 24 and 26), the group (BMI � 26)had a higher level of plasma 3-DG (both P < 0.05). In addition, asignificant difference of FBG was also found among three groups(5.04 � 0.47, 5.01 � 0.55, and 5.38 � 0.58, F = 4.060, P = 0.020) andthe group (BMI � 26) had a higher FBG than the other two groups(both P < 0.05), but no difference of PBG was found among threegroups (6.54 � 0.42, 6.53 � 0.51, and 5.73 � 0.48, F = 1.625,P = 0.201) (Fig. 1C).

Moreover, there was no correlation between the otherparameters (TG, TC, HDL-C AST, ALT, UA, and Scr) and plasma 3-DG level (r = �0.051, P = 0.561; r = 0.096, P = 0.277; r = �0.028,P = 0.750; r = �0.030, P = 0.732; r = �0.063, P = 0.472; r = 0.137,P = 0.118; r = 0.154, P = 0.0800 respectively).

3.3. The effect of elevated plasma 3-DG on glucose regulation

To observe the effects of elevated plasma 3-DG on glucoseregulation in seniors, we took a 2-year prospective study.According to the detected results (refer to ‘‘clinical examination’’for measurement design), the subjects were further sampled forinvestigation. The subjects with the BMI � 26, which had beenfound to be related with the increased level of 3-DG and bloodglucose (see above results), were excluded in the trial. As a result,we got 16 people with high level of 3-DG (�70 ng/ml) both atbaseline and 1 year later, and identified them as the continualhigh plasma 3-DG subjects (sampled for the study group). Thenwe randomly sampled 16 subjects (because of an absence, 14 infact) with the level of 3-DG less than 70 ng/ml both at baselineand 1 year later, and identified them as the normal plasma 3-DGsubjects (sampled for the control group). The plasma 3-DG level2 years after baseline of the study group was still higher thanthat of the control group (with ln(3-DG): 5.68 � 0.81 vs.

3.57 � 0.60, P < 0.01). Both age and BMI of the two groups showedno difference (with the age: 63.50 � 5.40 vs. 62.07 � 5.74, and BMI:24.61 � 3.25 vs. 25.13 � 2.27, both P > 0.05). All of these 30subjects got the OGTT, and HOMA-IR, ISI, HBCI, and DI60/DG60

were calculated.As shown in Fig. 2, compared with the control group, there were

higher level of FINs (10.51 � 4.84 vs. 6.81 � 4.02, P = 0.035), FBG(6.19 � 0.68 vs. 5.04 � 0.41, P = 0.004), and HOMA-IR (0.95 � 0.47 vs.

Fig. 2. The effect of elevated plasma 3-DG on glucose regulation. (A) The insulin of the two groups in OGTT: compared with the normal group, the continual high group had a

higher level of FINs (10.51 � 4.84 vs. 6.81 � 4.02, P = 0.035), a lower level of insulin at 1 h (70.01 � 51.64 vs. 73.27 � 47.50, P > 0.05) and 2 h (50.42 � 22.17 vs. 52.03 � 28.06,

P > 0.05). #Compared with normal group, P < 0.05. (B) The glucose of the two groups in OGTT: compared with the normal group, the continual high group had a higher level of FBG

(6.19 � 0.68 vs. 5.04 � 0.41, P = 0.004), a higher level of glucose at 1 h (10.01 � 4.87 vs. 9.65 � 2.77, P > 0.05), and at 2 h (9.41 � 5.27 vs. 7.66 � 2.59, P > 0.05). ##Compared with

normal group, P < 0.01. (C) The indexes of glycometabolism of the two groups: compared with the normal group, the continual high group had a higher HOMA-IR (0.95 � 0.47 vs.

0.29 � 0.54, P = 0.002), a lower ISI (�4.06 � 0.47 vs. �3.40 � 0.54, P = 0.002), and a lower insulin secrete index at 1 h during OGTT ([DI60/DG60]: 1.14 � 0.78 vs. 2.22 � 1.18,

P = 0.012). The HBCI of the two groups were similar (4.43 � 0.63 vs. 4.39 � 0.53, P > 0.05). Compared with normal group, #P < 0.05, ##P < 0.01.

G. Jiang et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 140–145 143

0.29 � 0.54, P = 0.002), but a lower level of ISI (�4.06 � 0.47 vs.

�3.40 � 0.54, P = 0.002) in the study group. It suggested that thecontinual high plasma 3-DG in those people might induce insulinresistance, and result in impaired fasting glucose with high level offasting insulin and glucose.

The HBCI of the study group was not smaller than the controlgroup (4.43 � 0.63 vs. 4.39 � 0.53, P > 0.05), which meant that thebasal insulin secretion of b-cell of the subjects with continual highplasma 3-DG was normal.

At 1 h during OGTT, compared with the control group, the studygroup had a slightly lower level of insulin (70.01 � 51.64 vs.

73.27 � 47.50, P > 0.05), a somewhat high level of glucose(10.01 � 4.87 vs. 9.65 � 2.77, P > 0.05), and a significant lowerDI60/DG60 when given IR correction ([DI60/DG60]: 1.14 � 0.78 vs.

2.22 � 1.18, P = 0.012). Then at 2 h during OGTT, the study group stillhad a slightly lower level of insulin (50.42 � 22.17 vs. 52.03 � 28.06,P > 0.05), and a somewhat higher level of glucose (9.41 � 5.27 vs.

7.66 � 2.59, P > 0.05). The results implied that there was adysfunction of b-cell to secrete insulin in response to high glucoseconcentration stimulation in the study group.

Table 2The incidence of impaired glucose regulation in the two groups.

Groups Impaired glucose regulation (n, %) (n) Incidence

(%)

NGT IGT IGT+IFG DM

Study group 6 (37.5%) 4 (25%) 4 (25%) 2 (12.5%) 16 62.5*

Control group 12 (85.71%) 2 (14.29%) 0 0 14 14.29

* Compared with control group, P < 0.05.

In addition, the study group had a high incidence rate ofimpaired glucose regulation (IGR) than the control group,x2 = 7.814, P = 0.029 (Table 2).

4. Discussion

On the basis of our previous studies [19], we have developeda reliable method with HPLC-FLD for quantification of 3-DG inhuman plasma. Our previous results showed that the plasma 3-DG concentration was less than 70 ng/ml in healthy youthbetween 23 and 30 years old (median: 20.47 ng/ml, range:13.41–61.79 ng/ml, n = 17), while more than 300 ng/ml inpatients with diabetes (median: 474.15 ng/ml, range: 71.74–1429.17 ng/ml, n = 25). It was also in agreement with theobservations described in others report [14]. Moreover, in thispaper, measures were planned and applied in the trial to ensureproper blood sampling for future assessment of 3DG (seemethods). On the other hand, endogenous 3-DG is known to besynthesized via the Maillard reaction [5], fructoseamine-3-kinase (F3K) [20], and the polyol pathway [21], and be detoxifiedto 3-deoxyfructose [22] and 2-keto-3-deoxygluconic acid [23].The plasma 3-DG levels may be elevated in some pathologicalstates, which could affect the pathways mentioned above, sopeople with some diseases such as hyperglycemia, nephrosis,and abnormal liver or renal function, were excluded toinvestigation. As a result, the FBG, PBG, AST, ALT, UA, and Scrall with normal levels showed no correlation with 3-DG in thesubjects. By using our method, in present study, we outlined forthe first time a skewed distribution of the plasma 3-DG levels inseniors (median: 43.52 ng/ml, range: 7.89–736.09 ng/ml). There

G. Jiang et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 140–145144

were about 35.6% of people with the plasma 3-DG level beyond70 ng/ml, and 10.61% of people over 300 ng/ml. Our resultsimplied that there was abnormal accumulation of plasma 3-DGin some elderly.

At present, the free radical theory of aging has been one ofthe modern hypotheses to explain the progressive deteriorationthat occurs in the aging organism [24]. Besides, glycation theory[25] and Maillard [26] were extended to explain the agingprocess. Carbonyl-protein cross-linking exists in both oxidationand glycation, which is suggested to be an important chemicalreaction that may be involved in lipofuscinogenesis [27].Glucose reacts non-enzymatically with protein amino groups,especially the long-lived proteins, such as collagen, then,spontaneously formed the highly reactive compounds. Somestudies had demonstrated that N

e-(carboxymethyl)lysine (CML),

methylglyoxal (MGO), or AGEs were detected in kidney [28],erythrocyte membrane [29], bone [30], and the crystalline lensof cataract [6] in elderly. It suggested that aging may also lead tothe elevated level of those products. In present study, we foundthat the 60–69 years group had a higher level of 3-DG than 50–59 years group (P < 0.001), and more importantly, a statisticallysignificant correlation between age and 3-DG level was foundamong the 50–66 years old people. In combination with resultsof our previous research, further analysis indicated that theplasma 3-DG level of senior was far higher than that of youth,but lower than that of diabetics. Furthermore, the averageplasma 3-DG level of people beyond 70 ng/ml was still far lowerthan that of diabetics. In addition, although we found thesubjects with BMI � 26 had a higher plasma 3-DG level, BMIshowed no differences among age-divided groups, and therewas also no correlation between BMI and age. In general,BMI dropped with aging [31]. So these findings not onlyfurther implied that increased plasma 3-DG might be involved inthe progress of aging, but also strongly demonstrated for thefirst time that aging maybe an independent factor of theformation and accumulation of endogenous 3-DG. In otherwords, 3-DG, a precursor of AGEs, was correlated with aging,just like AGEs.

Moreover, we consider the degeneration of the enzyme systemin elderly as one of the immediate causes accounting for theincrease of plasma 3-DG, since the ability to detoxify 3-DG by themetabolic enzyme system is important for maintaining the normallevel of 3-DG in physiological condition [21]. As to a little decreaseof plasma 3-DG level in the subjects beyond 70 years old, wededuced that it is mainly due to the decreasing content of proteinsand carbonyl compounds in the development of aging. The detailsremain to be elucidated.

In this paper, the association of BMI with plasma 3-DG wasalso found. According to the suitable cut-off of BMI for Chineseadult which is suggested by Cooperative Meta-analysis Group ofChina Obesity Task Force [12], and taking into account thedistribution characteristics of our samples, we divided thesubjects into three groups, and found that the group withBMI � 26 had a significantly higher level of 3-DG than the othertwo groups. It indicated that overweight or obesity wasassociated with the level of plasma 3-DG in elderly. As weknow, Maillard reaction between reducing sugars and aminoacids is a common reaction in foods which undergo thermalprocessing [32] and storage [33]. Foods processed by modernmethods are rich of sugar and protein that would have a largecontent of 3-DG, and would be absorbed [34]. Moreover, thiskind of diet is also one of the reasons for obesity. In addition,peroxidation of lipids can also lead to a multitude of reactivecarbonyl compounds [35]. The interrelationship amongthe overweight or obesity, plasma 3-DG level, and diets mayexplain in part why the subjects with BMI � 26 had a higher

plasma 3-DG level. In turn, it also may explain in part why theones beyond 70-year-old had the downtrend level of plasma 3-DG due to reduced food intake.

Although 3-DG may be a factor contributing to the disorder ofglucose regulation (described in preface), we thought it shouldbe a slow procedure requiring continual high level of plasma 3-DG in non-diabetics. To verify our hypothesis that the increasingaccumulation of plasma 3-DG may induce impaired glucoseregulation in seniors, which may eventually be involved ingenesis of senile diabetes, we applied a 2 years prospectivestudy. In the trial, to avoid the possible interferences, thesubjects investigated were further selected. There would be acorrelation between the level of blood glucose and age [36], aswell as the level of TG and TC [37], but we had not found thesame correlation in the subjects with normal level of FBG andPBG in our test. Obesity is not only a known risk factor fordiabetes, but is also associated with the level of plasma 3-DGand blood glucose observed in this paper, so the subjects withBMI � 26 were excluded. We investigated the effects ofcontinual high level of 3-DG versus normal level of 3-DG onglucose tolerance after 2 years follow-up, and found significant-ly elevated FBG, FINs, HOMA-IR, and decreased ISI, but the 1 hand 2 h blood glucose during OGTT showed no significantdifference in two groups. In our clinical study, the subjects hadtheir glucose tolerance impaired for at most 2 years, an earlystage of a disorder of glucose regulation, which may becharacterized as insulin resistance with high level of FBG andFINs mainly due to the decreased insulin sensitivity. It wasconsistent with the results reported newly [37]. Some studieshad reported that AGEs and MGO may inhibit glucose-stimulated insulin secretion both in vivo and in vitro

[10,38,39]. But in those studies, the subjects used were diabeticmodels, which had dysfunction of b-cell with absolute decreaseof insulin. In our study, our subjects with IGT or IFG were at theearly stage of a disorder, whose b-cells still had strongcompensatory secretion with normal HBCI. Nevertheless, theinsulin secrete index ([DI60/DG60]) of the group with continualhigh plasma 3-DG level was significantly smaller than thecontrol group when given IR correction. In combination with ourprevious findings [11], we attributed this to the insulinresistance of b-cell to explain the insufficiency of insulin at1 h and 2 h during OGTT. As the b-cell was not sensitive to thestimulation of elevated level of 1 h and 2 h glucose during OGTT,it failed to secrete sufficient insulin to reduce the postprandialglucose level. The results of the prospective study stronglysuggested that increasing accumulated plasma 3-DG couldlead to impaired glucose regulation in seniors. Decreasinginsulin sensitivity and reducing insulin secretion in thestimulation of high glucose concentration through inducinginsulin resistance of both peripheral tissues and b-cell are thepossible mechanism.

5. Conclusion

In present study, we outlined for the first time a skeweddistribution of the plasma 3-DG levels, and suggested that therewas abnormal elevation of plasma 3-DG in non-diabetic seniors ofSuzhou. Although there was some insufficiency (i.e. the smallsamples), it is sufficient to demonstrate that the increasingaccumulation of plasma 3-DG, which mainly resulted from aging,eventually lead to the impaired glucose regulation throughdecreasing insulin sensitivity and reducing insulin secretion,inducing insulin resistance of both peripheral tissues and b-cellis the possible mechanism. These results further imply that 3-DG isone of risk factors for diabetes in elderly, especially the elderly withlarge BMI.

G. Jiang et al. / Diabetes & Metabolic Syndrome: Clinical Research & Reviews 6 (2012) 140–145 145

Conflict of interest

None.

Acknowledgments

This study was supported by the research funds from JiangsuProvince Administration of Traditional Chinese Medicine(LZ11145), and the research fund from Suzhou Science andTechnology Department (SZD09139).

References

[1] Sakiyama H, Takahashi M, Yamamoto T, Teshima T, Lee SH, Miyamoto Y, et al.The internalization and metabolism of 3-deoxyglucosone in human umbilicalvein endothelial cells. Journal of Biochemistry 2006;139:245–53.

[2] Kiho T, Asahi T, Usui S, Matsunaga T, Ukai S. Effect of 3-deoxyglucosone on theactivities of enzymes responsible for glucose metabolism in mouse liver.Biological and Pharmaceutical Bulletin 1996;19:1106–8.

[3] Eriksson UJ, Wentzel P, Minhas HS, Thornalley PJ. Teratogenicity of 3-deox-yglucosone and diabetic embryopathy. Diabetes 1998;47:1960–6.

[4] Kusunoki H, Miyata S, Ohara T, Liu BF, Uriuhara A, Kojima H, et al. Relationbetween serum 3-deoxyglucosone and development of diabetic microangio-pathy. Diabetes Care 2003;26:1889–94.

[5] Niwa T, Takeda N, Miyazaki T, Yoshizumi H, Tatematsu A, Maeda K, et al.Elevated serum levels of 3-deoxyglucosone, a potent protein-cross-linkingintermediate of the Maillard reaction, in uremic patients. Nephron 1995;69:438–43.

[6] Gul A, Rahman MA, Salim A, Simjee SU. Advanced glycation end products insenile diabetic and nondiabetic patients with cataract. Journal of Diabetes andits Complications 2009;23:343–8.

[7] Yaffe K, Lindquist K, Schwartz AV, Vitartas C, Vittinghoff E, Satterfield S, et al.Advanced glycation end product level, diabetes, and accelerated cognitiveaging. Neurology 2011;77:1351–6.

[8] Basta G, Schmidt AM, De Caterina R. Advanced glycation end products andvascular inflammation: implications for accelerated atherosclerosis in diabe-tes. Cardiovascular Research 2004;63:582–92.

[9] Ankrah NA, Appiah-Opong R. Toxicity of low levels of methylglyoxal: depletionof blood glutathione and adverse effect on glucose tolerance in mice. Toxicol-ogy Letters 1999;109:61–7.

[10] Zhao Z, Zhao C, Zhang XH, Zheng F, Cai W, Vlassara H, et al. Advanced glycationend products inhibit glucose-stimulated insulin secretion through nitric ox-ide-dependent inhibition of cytochrome c oxidase and adenosine triphosphatesynthesis. Endocrinology 2009;150:2569–76.

[11] Wang Q, Jiang Gr, Zhang Lr. Effects of 3-deoxyglucosone on blood glucose ofnormal mice. Chinese Journal of Diabetes 2010;18:220–2.

[12] Zhou B. Predictive values of body mass index and waist circumference to riskfactors of related diseases in Chinese adult population. Zhonghua Liu Xing BingXue Za Zhi 2002;23:5–10.

[13] Haffner SM, Kennedy E, Gonzalez C, Stern MP, Miettinen H. A prospectiveanalysis of the HOMA model. The Mexico City Diabetes Study. Diabetes Care1996;19:1138–41.

[14] Hamada Y, Nakamura J, Fujisawa H, Yago H, Nakashima E, Koh N, et al. Effectsof glycemic control on plasma 3-deoxyglucosone levels in NIDDM patients.Diabetes Care 1997;20:1466–9.

[15] Haffner SM, Miettinen H, Stern MP. The Homeostasis Model in the San AntonioHeart Study. Diabetes Care 1997;20:1087–92.

[16] Li Guangwei, Hu Yinghua, Yang Wenging, Jiang Yayun, Cao Huibi, WangJinping, et al. Effects of insulin resistance and insulin secretion on the devel-opment of type II diabetes mellitus. Chinese Journal of Internal Medicine1998;9:600–4.

[17] Organization WH. Definition, diagnosis and classification of diabetes mellitusand its complications. Report of a WHO Consultation. WHO/NCD/NCS/992.Geneva: WHO; 1999.

[18] Hasuike Y, Nakanishi T, Otaki Y, Nanami M, Tanimoto T, Taniguchi N, et al.Plasma 3-deoxyglucosone elevation in chronic renal failure is associated withincreased aldose reductase in erythrocytes. American Journal of Kidney Dis-eases 2002;40:464–71.

[19] Jiang Gr, Zhang Lr. Analysis of 3-deoxyglucosone by HPLC and preliminaryapplication in clinical sample. Chinese Journal of Hospital Pharmacy 2008;802–4.

[20] Senior K. The end for diabetic kidney disease? Drug Discovery Today 2001;6:601–2.

[21] Liang ZQ, Hayase F, Kato H. Purification and characterization of NADPH-dependent 2-oxoaldehyde reductase from porcine liver. A self-defence en-zyme preventing the advanced stage of the Maillard reaction. EuropeanJournal of Biochemistry 1991;197:373–9.

[22] Knecht KJ, Feather MS, Baynes JW. Detection of 3-deoxyfructose and 3-deoxyglucosone in human urine and plasma: evidence for intermediate stagesof the Maillard reaction in vivo. Archives of Biochemistry and Biophysics1992;294:130–7.

[23] Fujii E, Iwase H, Ishii-Karakasa I, Yajima Y, Hotta K. The presence of 2-keto-3-deoxygluconic acid and oxoaldehyde dehydrogenase activity in human ery-throcytes. Biochemical and Biophysical Research Communications 1995;210:852–7.

[24] Harman D, Aging:. a theory based on free radical and radiation chemistry.Journal of Gerontology 1956;11:298–300.

[25] Cerami A, Hypothesis. Glucose as a mediator of aging. Journal of the AmericanGeriatrics Society 1985;33:626–34.

[26] Monnier VM, Cerami A. Nonenzymatic browning in vivo: possible process foraging of long-lived proteins. Science 1981;211:491–3.

[27] Yin DZ. Lipofuscin-like fluorophores can result from reactions betweenoxidized ascorbic acid and glutamine. Carbonyl-protein cross-linkingmay represent a common reaction in oxygen radical and glycosylation-related ageing processes. Mechanisms of Ageing and Development 1992;62:35–45.

[28] Semba RD, Fink JC, Sun K, Bandinelli S, Guralnik JM, Ferrucci L. Carboxymethyl-lysine, an advanced glycation end product, and decline of renal function inolder community-dwelling adults. European Journal of Nutrition 2009;48:38–44.

[29] Jha R, Rizvi SI. Carbonyl formation in erythrocyte membrane proteins duringaging in humans. Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub2011;155:39–42.

[30] Tang SY, Zeenath U, Vashishth D. Effects of non-enzymatic glycation oncancellous bone fragility. Bone 2007;40:1144–51.

[31] Fan C, Wang CX, Huang YF. The investigation and analysis of body mass indexand hypertension in 2841 retirees of Suzhou urban population. ChineseJournal of Gerontology 2010;30:2821–3.

[32] Jaeger H, Janositz A, Knorr D. The Maillard reaction and its control during foodprocessing. The potential of emerging technologies. Pathologie Biologie2010;58:207–13.

[33] Babsky NE, Toribio JL, Lozano JE. Influence of storage on the composition ofclarified apple juice concentrate. Journal of Food Science 1986;51:564–7.

[34] Kato H, van Chuyen N, Shinoda T, Sekiya F, Hayase F. Metabolism of 3-deoxyglucosone, an intermediate compound in the Maillard reaction, admin-istered orally or intravenously to rats. Biochimica et Biophysica Acta1990;1035:71–6.

[35] Baynes JW, Thorpe SR. Glycoxidation and lipoxidation in atherogenesis. FreeRadical Biology and Medicine 2000;28:1708–16.

[36] Singh GM, Danaei G, Pelizzari PM, Lin JK, Cowan MJ, Stevens GA, et al. The ageassociations of blood pressure, cholesterol, and glucose/clinical perspective.Circulation 2012;125:2204–11.

[37] Wang W, Zhao D, Wu Z, Wu G, Liu J, Zeng Z, et al. Study of triglyceridedistribution aged 35–64, and their association with other cardiovasculardisease risk factors in 11 provinces. Zhonghua Liu Xing Bing Xue Za Zhi2001;22:26–9.

[38] Coughlan MT, Yap FY, Tong DC, Andrikopoulos S, Gasser A, Thallas-Bonke V,et al. Advanced glycation end products are direct modulators of beta-cellfunction. Diabetes 2011;60:2523–32.

[39] Riboulet-Chavey A, Pierron A, Durand I, Murdaca J, Giudicelli J, Van ObberghenE. Methylglyoxal impairs the insulin signaling pathways independently of theformation of intracellular reactive oxygen species. Diabetes 2006;55:1289–99.