点击显示 收起
1 From the Institute of Brain Chemistry and Human Nutrition, London Metropolitan University, London, United Kingdom (YM, KG, BT, and MC), and the Endocrine and Diabetic Day Centre, Guy's and St Thomas' Hospital Trust, London, United Kingdom (CL and BO-S)
2 Supported by Diabetes UK, the March of Dimes, the President Club of The Mother and Child Foundation, and Guy's and St Thomas' Charity (Endocrine and Diabetic Fund). 3 Reprints not available. Address correspondence to Y Min, Institute of Brain Chemistry and Human Nutrition, London Metropolitan University, 166-220 Holloway Road, London N7 8DB, UK. E-mail: y.min{at}londonmet.ac.uk.
ABSTRACT
Background: Pregestational maternal diabetes increases obesity and diabetes risks in the offspring. Both conditions are characterized by insulin resistance, and diabetes is associated with low membrane arachidonic (AA) and docosahexaenoic (DHA) acids.
Objective: We investigated whether type 1 and type 2 diabetes in pregnancy compromise maternal and fetal membrane essential fatty acids (FAs).
Design: We studied 39 nondiabetic (control subjects), 32 type 1 diabetic, and 17 type 2 diabetic pregnant women and the infants they delivered. Maternal and cord blood samples were obtained at midgestation and at delivery, respectively. Plasma triacylglycerols and choline phosphoglycerides and red blood cell (RBC) choline and ethanolamine phosphoglyceride FAs were assessed.
Results: The difference in maternal plasma triacylglycerol FAs between groups was not significant. However, the type 1 diabetes group had lower plasma choline phosphoglyceride DHA (3.7 ± 0.9%; P < 0.01) than did the control group (5.2 ± 1.6%). Likewise, RBC DHA was lower in the type 1 [choline: 3.4 ± 1.5% (P < 0.01); ethanolamine: 5.9 ± 2.5% (P < 0.05)] and type 2 [choline: 3.5 ± 1.6% (P < 0.05)] diabetes groups than in the control group (choline: 5.5 ± 2.2%; ethanolamine: 7.5 ± 2.5%). Cord AA and DHA were lower in the plasma (type 1: P < 0.01) and RBC (type 2: P < 0.05) choline phosphoglycerides of the diabetics than of the control subjects, and cord RBC ethanolamine phosphoglycerides were lower in DHA (P < 0.05) in both diabetes groups than in the control group.
Conclusions: Diabetes (either type) compromises maternal RBC DHA and cord plasma and RBC AA and DHA. The association of these 2 FAs with insulin sensitivity may mean that the current finding explains the higher incidence of insulin resistance and diabetes in the offspring of diabetic women.
Key Words: Type 1 diabetes • type 2 diabetes • pregnancy • fetus • arachidonic acid • docosahexaenoic acid
INTRODUCTION
The rapidly growing incidence of childhood obesity and type 2 diabetes has become a global concern. Evidence indicates that intrauterine exposure to type 1 or type 2 diabetes increases the risk of obesity, insulin resistance, insulin secretory defect, and subsequent development of type 2 diabetes in the offspring (1-5). In addition, evidence is good that prenatal and postnatal nutrition plays a pivotal role in the child's predisposition to these conditions (6-8).
Lipid metabolism abnormalities are well described in type 1 and type 2 diabetes. One of the characteristics is low concentrations of arachidonic acid (AA) or docosahexaenoic acid (DHA) (or both) in the circulating and tissue phospholipids (9-13). AA and DHA are the major metabolites of essential fatty acids (FAs)—namely, linoleic acid (the n–6 family) and -linolenic acid (the n–3 family), respectively. They are essential structural and functional components of organs such as blood vessels, pancreas ß cells, retina, and brain (14-18). Changes in the concentrations of AA and DHA in the muscle and ß-cell phospholipids exert an influence on insulin sensitivity (19) and glucose and insulin concentrations (6, 20). Moreover, AA is thought to act as a second messenger in insulin secretion (21, 22).
Previous studies showed that gestational diabetes is associated with low AA and DHA in the red blood cells (RBCs) of pregnant women (23) and in the plasma and RBCs of their neonates (24, 25). There are few such studies on type 1 and type 2 diabetes (26-28). Moreover, those studies either were based on a small number of subjects, analyzed only total phospholipids FAs (26, 27), or did not test for the effect on membrane FAs (28). In the current study, we investigated the effect of pregestational diabetes (type 1 and type 2) during pregnancy on the plasma and RBC FAs of the mothers at midgestation (between 25 and 30 wk) when maternal fat accumulation reaches its maximum (29) and of the fetus (cord blood) when the demand for AA and DHA is high (30-32).
SUBJECTS AND METHODS
Subjects
Expectant mothers with pregestational type 1 and type 2 diabetes and those without diabetes (control subjects) were recruited during the first trimester. The exclusion criteria were twin pregnancy, preexisting high blood pressure, preeclampsia in previous pregnancy, other chronic disorders, habitual use of illegal drugs, a positive test for gestational diabetes, underage pregnancy (<16 y old), and no cesarean section. The control subjects were screened for gestational diabetes between the 28th and 34th weeks of pregnancy with the use of an oral-glucose-tolerance test (33). After an overnight fast, the women were given a 75-g glucose load, and their blood glucose concentration was monitored at 60 and 120 min. If the blood glucose concentration at 60 min was > 8 mmol/L, a second sample was taken at 120 min. Gestational diabetes was diagnosed if the second reading was 9 mmol/L.
During pregnancy, the women with type 1 and 2 diabetes were treated with insulin, either 2 times/d short and intermediate preparations or a 4 times/d basal bolus regimen with each meal and intermediate-acting insulin at bedtime. The insulin dose was self-adjusted on the basis of the results of the blood glucose measurements and at the 2-wk interval consultation with a diabetologist at the combined antenatal and diabetic clinic.
Infants born prematurely (before 37 complete weeks of gestation) or at low birth weight (< 2.5 kg) were excluded from the study.
Written informed consent was obtained from all participating women. Ethical approval was granted by the ethics committee of the Guy's and St Thomas' Hospital Trust. This investigation was carried out in accordance with the principles of the Declaration of Helsinki as revised in 2000.
Blood sample collection
This investigation was carried out at the antenatal clinic and diabetic endocrine day center of St Thomas' Hospital in London. Fasting venous blood samples (5 mL) were collected from the mothers at midgestation (25–30 wk) and from cord blood at delivery into heparin-containing tubes. Blood was also collected from the diabetic women during the first (13 wk) and third (26 wk) trimesters as part of a routine clinical management and glycemic monitoring. RBCs were separated from plasma by cold centrifugation at 3000 rpm for 15 min at 4 °C (CR3i; Jouan, Ilkeston, United Kingdom) and washed twice with an equal volume of 0.85% saline (NaCl). The samples were subsequently transported to the laboratory and stored at –70 °C until they were analyzed.
Sample analysis
Plasma and RBC total lipids were extracted by using chloroform and methanol (34). Phospholipids were separated by thin-layer chromatography on silica gel plates with the use of the developing solvents chloroform:methanol:water at a ratio of 60:30:4 (by volume). Neutral lipid was recovered from the plasma phospholipid plate, and triacylglycerols were obtained with the use of a developing solvent system, petrol spirit:ether:formic acid:methanol at a ratio of 85:15:2.5:1 (by volume). We used 0.01% 2,6-di-tert-butyl-p-cresol (butylated hydroxy toluene) in all solvents throughout to prevent potential oxidation. FA methyl esters were prepared in 15% acetyl chloride in dry methanol solution at 70 °C for 3 h and separated by using a gas-liquid chromatograph (HRGC MEGA 2 Series; Fisons Instruments, Milan, Italy) fitted with a BP20 capillary column (25 m x 0.32 mm internal diameter, 0.25-µm film; Fisons Instrumentss). Hydrogen was used as a carrier gas, and the injector, oven, and detector temperatures were 250, 200, and 280 °C, respectively. The FA methyl esters were identified by a comparison of retention time with authentic standards. Peak areas were quantified by using a computer chromatography data system (EZChrom Chromatography Data System; Scientific Software, San Ramon, CA).
Statistical analysis
Differences in demographic values and FAs between 3 study groups (healthy control subjects and type 1 and type 2 diabetic subjects) were tested by using one-way analysis of variance with Bonferroni adjustment for multiple comparisons. Relations between variables were examined by using the Pearson correlation coefficient. Data are presented as means (±SDs) of total FAs (%) unless stated otherwise, and significance was set at P < 0.05. SPSS for WINDOWS statistical software (version 11.5.0; SPSS Inc, Chicago, IL) was used in all analyses.
RESULTS
A total of 39 control subjects, 32 type 1 diabetic women, and 17 type 2 diabetic women were studied. Demographic and clinical characteristics of subjects are shown in Table 1. The mean duration of the disease before pregnancy was 17.9 y in the type 1 and 2.1 y in the type 2 diabetic subjects. The mean prepregnancy body mass index (BMI; in kg/m2) of the diabetic women was within the overweight range (BMI 25), and that of the controls within the normal-weight range (35). Five infants (n = 4 and 1 for type 1 and type 2 diabetes, respectively) were born prematurely, and 1 infant of a type 1 diabetic woman had low birth weight. Of the type 2 diabetes group, one woman had a macrosomic infant, and another developed preeclampsia. One of the women in the type 1 diabetes group was diagnosed with pregnancy-induced hypertension. The FA data for these women were included in the analysis. Cord blood was not available from 6 control subjects or from 1 woman with type 1 and 1 woman with type 2 diabetes. The mean gestational age of the infants of the healthy controls was greater than that of the infants of the women in the type 1 and type 2 diabetes groups (P < 0.0001). The data analyses were based on 33 infants of control subjects, 26 infants of type 1 diabetes subjects, and 15 infants of type 2 diabetes subjects.
View this table:
TABLE 1. Characteristics of the study participants1
Maternal fatty acids
As shown in Table 2, there was no difference in the concentrations of the major saturated, monounsaturated, n–6, and n–3 FAs or Mead acid in plasma triacylglycerols between the control subjects and the diabetic women. Likewise, the FA composition of the plasma choline phosphoglycerides of the women with type 1 and type 2 diabetes did not differ significantly from that of the control subjects, except DHA, which was significantly lower in the women with type 1 diabetes (P < 0.01).
View this table:
TABLE 2. Fatty acid (FA) composition of maternal plasma triacylglycerols and choline phosphoglycerides1
RBC choline phosphoglycerides and ethanolamine phosphoglycerides represent the prominent phospholipids of the outer and inner cell membranes, respectively. The FA composition of RBCs is shown in Table 3. In contrast to the plasma choline phosphoglycerides, the concentration of palmitic acid was significantly higher in both diabetes groups (P < 0.05) than in the control group. Moreover, the RBC choline phosphoglycerides of the women with type 1 diabetes contained higher concentrations of oleic acid (P < 0.01) and total saturates and monoenes (P < 0.01) and lower concentrations of total n–6 (P < 0.05) and n–3 (P < 0.01) than did those of the control subjects. The proportion of DHA was lower in RBC choline phosphoglycerides of the women in with type 1 (–38%, P < 0.01) and type 2 (–36%, P < 0.05) diabetes than in those of the control subjects. Moreover, the RBC ethanolamine phosphoglyceride DHA was lower by 21% in the women with type 1 diabetes than in the control subjects (P < 0.05).
View this table:
TABLE 3. Fatty acid (FA) composition of maternal red blood cell choline and ethanolamine phosphoglycerides1
Cord blood (fetal) fatty acids
The FA composition of the cord plasma triacylglycerols and choline phosphoglycerides and the RBC choline phosphoglycerides and ethanolamine phosphoglycerides are summarized in Table 4 and Table 5, respectively. The concentrations of the major triacylglycerol saturated and monounsaturated FAs in the cord blood of the infants of the women with type 1 diabetes did not differ from those in the cord blood of infants of the control subjects. Similarly, the concentrations of the parent n–6 FA, linoleic acid, and of its major metabolite, AA, in the cord blood did not differ significantly between the type 1 diabetic infants and the control infants. However, the cord blood of type 1 diabetic infants contained lower concentrations of -linolenic acid (P < 0.01) and its C22 metabolites (P < 0.05) than in that of the control infants. In contrast, the triacylglycerol FA composition of the cord blood of the type 2 diabetic infants did not differ significantly from that of the control subjects.
View this table:
TABLE 4. Fatty acid (FA) composition of cord plasma triacylglycerols and choline phosphoglycerides1
View this table:
TABLE 5. Fatty acid (FA) composition of cord red blood cell choline and ethanolamine phosphoglycerides1
In the cord blood of type 1 diabetic infants, the plasma choline phosphoglycerides contained higher proportions of palmitic (P < 0.01), palmitoleic (P < 0.01), and oleic (P < 0.05) acids and lower proportions of AA and DHA (P < 0.01) and total n–6 and n–3 (P < 0.05) than did those of the control infants. The cord blood of type 2 diabetic infants contained lower concentrations of AA (P < 0.05) and higher concentrations of palmitoleic (P < 0.05) and linoleic (P < 0.01) acids in the plasma choline phosphoglycerides than did the cord blood of the control infants. However, the saturated and n–3 FAs in the cord blood of type 2 diabetic infants did not differ significantly from that of the control infants.
In RBC choline phosphoglycerides of the cord blood, there was no difference between type 1 diabetic and control subjects, except with respect to palmitoleic and Mead acids, which were higher in the type 1 diabetic subjects (P < 0.05). The cord blood of the type 2 diabetic infants had higher concentrations of palmitic and oleic acids (P < 0.01), total saturates (P < 0.05), and monoenes (P < 0.01), whereas the concentrations of stearic acid (P < 0.01), AA (P < 0.01), DHA (P < 0.05), and total n–6 and n–3 (P < 0.05) were lower than in the control infants. The cord blood of both types of diabetics contained higher oleic acid and total monoenes (P < 0.05) but lower DHA (P < 0.05) in RBC ethanolamine phosphoglycerides than did the cord blood of the control subjects.
DISCUSSION
RBC membrane FA abnormality was evident in both the type 1 and type 2 diabetic women and their neonates at birth. The abnormality in the mothers could be a reflection of low FA intake or the synergistic effect of the disease and pregnancy. Moore et al (36) reported that dietary linoleic acid correlates with plasma or serum triacylglycerol linoleic acid concentrations. Triacylglycerol linoleic acid concentrations did not differ significantly between the diabetic and control women. Hence, diet could not have been the factor for the discrepancy in plasma and RBC phospholipid FAs. The activity of the enzymes -6 and -5 desaturase, which are vital for the synthesis of AA and DHA, has been shown to be impaired in uncontrolled diabetes, leading to a decrease in their proportional contribution to plasma and tissue membrane phospholipids. Both type 1 and type 2 diabetic women achieved reasonable glycemic control, so, if the degree of glycemic control determines phospholipid FA composition, then there may be a correlation with glucose or glycated hemoglobin (HbA1c). However, no correlation was found between blood glucose concentrations or HbA1c and RBC AA and DHA concentrations. Moreover, neither of the indirect markers of desaturase activity (ie, ratios of 20:3n–6 to 18:2n–6 for -6 and of 20:4n–6 to 20:3n–6 for -5) in diabetic women differed significantly from those in control subjects. Yet, consistent with our previous finding in neonates of type 1 diabetic women (28), the proportion of AA and DHA in plasma choline phosphoglycerides of the subjects with type 1 diabetes was lower than that in the control subjedts. This suggests, as in experimental model of diabetes (37-39), that insulin treatment may not fully correct plasma and RBC FA composition in type 1 human diabetes.
It is interesting that the degree of FA abnormality was less pronounced in the type 2 than in the type 1 diabetic group; the effect was mainly on RBC phospholipid DHA. Studies in type 2 diabetes have yielded inconsistent results. Some studies found no changes (40, 41), whereas others reported an increase (42, 43) or reduction (44, 45) in n–6 and n–3 long-chain FAs. This discrepancy could be due to differences in study subjects (ie, sex, age, duration of diabetes, treatment, and presence of diabetes-related complications) or to the analysis of different lipid fractions. Nonetheless, low membrane DHA is a common feature of insulin-resistant conditions (11, 19, 46, 47).
By the early third trimester, maternal n–6 and n–3 fat accretion maximizes and declines progressively, which parallels the rapid accumulation of AA and DHA in the fetus (29-32). Because fetal essential FA status is directly influenced by maternal status, one can speculate that a low maternal AA and DHA status might account for the reduced concentrations of these 2 FAs in the fetuses of the diabetic women. However, previous studies showed that diabetes in pregnancy exaggerates not only maternal hyperlipidemia (48, 49) but also placental lipoprotein lipase activity and expression of FA-binding proteins that display high affinity for long-chain polyunsaturated FAs (50, 51; T Radaelli, A Varastehpour, P Catalano, S Hauguel-De Mouzon, unpublished observations, 2003). Thus it is conceivable that the mobilization or release (or both) of these particular FAs from the placenta may be compromised in diabetes, which would diminish the transfer of the FA to the fetal circulation. Although maternal diabetes was likely to be mainly responsible for the abnormal FA status observed in the fetuses in the current study, genetic effects cannot be ruled out. Indeed, Kunesova et al (52) suggested that there is a strong genetic influence over a specific FA for membrane composition.
Alternatively, the abnormal fetal FAs could be due to altered lipid metabolism in itself in the presence of fetal hyperglycemia and hyperinsulinemia. Whereas patients with type 2 diabetes generally are insulin resistant, patients with type 1 diabetes primarily are insulin sensitive. Nevertheless, both groups have inappropriately high plasma glucose concentrations, as do their fetuses. Thus, like the mothers' tissue, the fetal tissues would have been exposed to an inappropriate plasma glucose:insulin. Could this metabolic perturbation be responsible for the fetal RBC membrane FA abnormality, or is it merely a marker of insulin resistance?
AA and DHA are vital to the structure and function of all cell membranes, including RBCs. Consequently, a loss of these FAs is likely to affect the membrane's conformation and function (53, 54). In humans, the concentration of DHA in cell membranes has been linked to insulin sensitivity (the lower the membrane DHA, the greater the insulin resistance). It has also been shown that supplementation of experimental animals with AA improves insulin sensitivity by enhancing glucose uptake (55, 56). Hence, our results may provide a possible explanation for the higher incidence of insulin resistance and diabetes in the offspring of diabetic women than in those of the control subjects.
ACKNOWLEDGMENTS
We thank the mothers for their cooperation.
YM participated in the study design, conducted lipid analysis, performed data analysis, and wrote the manuscript. CL, KG, and MC participated in the study design and provided critical review of the manuscript. CL and KG assisted with the editing of the manuscript. BT was involved in the lipid analysis, and BO was responsible for recruiting patients and collecting samples. All authors participated in the analytical discussion of the results and approved the final version of the manuscript. None of the authors had any personal or financial conflict of interest.
REFERENCES