Research Article | Open Access
Effects of Lipoic Acid Supplementation on Activities of Cyclooxygenases and Levels of Prostaglandins E2 and F2α Metabolites, in the Offspring of Rats with Streptozotocin-Induced Diabetes
Background. Our aim was to evaluate the protective effect of lipoic acid (LA) on fetal outcome of diabetic mothers. Methods. Diabetes was induced in female rats using streptozotocin and rats were made pregnant. Pregnant control (group 1; ; and group 2; ) or pregnant diabetic (group 3; ; and group 4; ) rats were treated daily with either LA (groups 2 and 4) or vehicle (groups 1 and 3) between gestational days 0 and 15. On day 15 of gestation, the fetuses, placentas, and membranes were dissected, examined morphologically, and then homogenized, to measure cyclooxygenase (COX) activities and metabolisms of prostaglandin (PG) E2 (PGEM) and PGF2α (PGFM) levels. The level of total glutathione was measured in the maternal liver and plasma and in all fetuses. Results. Supplementation of diabetic rats with LA was found to significantly () reduce resorption rates in diabetic rats and led to a significant () increase in liver, plasma, and fetuses total glutathione from LA-TD rats as compared to those from V-TD. Decreased levels of PGEM and elevated levels of PGFM in the fetuses, placentas, and membranes were characteristic of experimental diabetic gestation associated with malformation. The levels of PGEM in malformed fetuses from LA-TD mothers was significantly () higher than those in malformed fetuses from V-TD rats. Conclusions. LA treatment did not completely prevent the occurrence of malformations. Thus, other factors may be involved in the pathogenesis of the diabetes-induced congenital malformations.
Mothers suffering from diabetes mellitus have high chances of loss of pregnancy and giving birth to babies with congenital malformations. These congenital malformations most commonly involve the central nervous, cardiovascular, and skeletal systems . There has been an ongoing research that seeks to identify the factors that have been involved in diabetic embryopathy, including increased levels of glucose and ketone bodies [2, 3] and reduced levels of arachidonic acid and prostaglandin E2 (PGE2) [4, 5]. Furthermore, several studies suggested that disequilibrium of oxidant-antioxidants levels prevalent in diabetic patients is a potential cause of these aberrations [6–11]. Increased formation of reactive oxygen species in association with increased teratogenesis has been demonstrated in rat embryos cultured in vitro in medium containing high levels of glucose . The addition of antioxidants and superoxide dismutase to the medium decreased the rate of embryonic dysmorphogenesis, which suggests the involvement of reactive oxygen species in the teratogenic process . In vivo studies using several models of diabetes suggested that administration of vitamin E , vitamin C , and lipoic acid  was found to be useful in improving fetal outcome in in vivo models of diabetic embryopathy.
Lipoic acid (LA, also known as alpha-lipoic acid or thioctic acid) is found inside every cell of the body. Lipoic acid is a potent antioxidant, capable of inducing its activity in both a lipophilic and an aqueous milieu [12, 13]. It also helps to regenerate both fat and water soluble antioxidant vitamins  and improve sugar metabolism and energy production [15, 16]. Furthermore, α-LA was shown to reverse apoptosis, ameliorate mitochondrial deformation, and increase the amount of mitochondrial DNA .
A healthy body makes enough LA to supply its energy requirements; therefore, there is no daily requirement for this supplement. However, several medical conditions appear to be accompanied by low levels of LA—specifically, diabetes, liver cirrhosis, and heart disease —which suggests that supplementation would be helpful.
At present, the exact mode of action of LA has not been fully elucidated. This research study is designed to investigate the protective nature of LA on fetal outcome while determining the processes involved in diabetic fetopathy. To achieve this, the study will examine the effects of LA on the activity of cyclooxygenases and the levels of prostaglandins E2 and F2α metabolites in diabetic rats.
This study will make major contribution to our nation as it will reduce the rate of congenital malformation and will be of considerable benefit to health sector since congenital malformation imposes a substantial cost to the health budget and emotional burdens upon society.
2.1. Experimental Animals
All animal studies were reviewed and approved by the Animal Ethics Committee of University of Petra (no. ANIM: LAGES-2014-04). The guide for the care and use of laboratory animals was followed.
Nulliparous female rats from a local Wistar-derived strain, 60 to 70 days of age with initial weights of 220 ± 30 g, were used. They were housed in rooms maintained at a temperature of 21°C ± 3°C and 12 : 12 h light/dark cycle. A commercial diet and tap water were provided ad libitum.
2.2. Experimental Diabetes Induction
The rats were randomly distributed to four groups, two control groups (pregnant nondiabetic; groups 1 and 2) and two diabetic groups (groups 3 and 4). Diabetes was induced in female rats from groups 3 (n = 10) and 4 (n = 8) by a single intraperitoneal (IP) injection of freshly prepared streptozotocin (STZ; Sigma-Aldrich Chemical Co., Poole, UK; 65 mg/kg, dissolved in citrate buffer, 100 mmol/L; pH 4.5), while control animals groups 1 (n = 9) and 2 (n = 7) were injected with a buffer solution (an equivalent volume of STZ injection). One week after the injections, venous tail blood was collected and hyperglycemia was confirmed with blood glucose levels greater than 11 mmol/L using a glucometer.
The rats (both controls and diabetic) were mated with healthy nondiabetic males overnight. The presence of a sperm mucus plug in the vagina the following morning signified pregnancy. Further, the presence of spermatozoa was designated as day zero of pregnancy. Pregnancy was confirmed by subsequent increases in body weight.
2.4. Lipoic Acid Induction
Thereafter, rats were daily injected IP with either LA (30 mg/kg body weight in Tris buffer with pH 7.4; groups 2 and 4) or vehicle only (groups 1 and 3) between gestational days 0 and 15.
2.5. Sample Collection
At day 15 of gestation, pregnant diabetic and control rats were sacrificed by cervical dislocation, and maternal blood samples were collected by cardiac puncture. In each sacrificed rat, the uterus was exposed by cesarean section, and the number of implantation and resorption processes was recorded. The live fetuses, along with their placentas and membranes, were dissected out of the uterine horns, rinsed carefully in phosphate buffered saline (PBS), and carefully labeled. Overall growth and differentiation of the fetuses were quantified by direct measurement of crown-to-rump length (CRL). The fetuses were examined for their general morphology, and the presence of any disturbed fetal development, such as an open neural tube or growth retardation, and the appearance of specific parts and organs (head, ear, legs, tail, and body rotation) were also noted. This examination was conducted by a researcher masked to the various experimental groups. Fetuses with apparent anomalies in one or more of these structures were considered “malformed.” Fetal, placental, and membrane weights were also recorded. The fetuses, placentas, and membranes were then snap-frozen in liquid nitrogen and kept at −80°C for subsequent analysis.
2.6. Sample Preparation and Assays
To rule out the possible presence of COX-1 inhibitor in association with pregnancy, we examined the activity of COX-1 in vitro by the addition of supernatants from the homogenates of placentas, which could have contained such an inhibitor, to the incubated media, and found that such activity was not affected.
The COX activity assays were determined according to procedures used previously . Protein was determined using a modification of Bradford’s method . The specific activity of each enzyme was determined by dividing its activity by the protein concentration in the sample (U/mg protein). The intra- and interassay coefficients of variation for COX were 2.6% and 5.4%, respectively.
Circulating prostaglandin (PG) PGF2α and PGE2 are rapidly converted in vivo to their 13,14-dihydro-15-keto metabolites PGFM and PGEM, respectively, [20, 21] with more than 90% cleared from circulation by a single passage through the lungs. Although levels of the enzymes that metabolize prostaglandins vary in different tissues, it is considered that measurement of their metabolites is used to provide a reliable estimate of actual PGF2α and PGE2.
PGFM and PGEM were measured by enzyme immunoassays (EIA)  with a sensitivity of 11 pg/mL and 120 pg/mL, respectively. The intra- and interassay coefficients of variation for PGFM were 10% and 21% and for PGEM were 8% and 8%, respectively.
Biochemical parameters were determined for all individual fetuses and placentas, while membranes from each group were pooled to allow determination of all parameters in the same sample. PGs metabolites were also assayed in maternal blood.
Liver samples from the maternal rats were removed, then snap-frozen in liquid nitrogen, and kept at −80°C for subsequent analysis. Liver and plasma total glutathione levels, a sensitive indicator of oxidative stress, were determined and expressed as μmol/g tissue and μmol, respectively. The level of total glutathione was measured also in the fetuses of all groups and expressed as μmol/mg of protein. The kits for COX, EIA for both PGFM and PGEM, and total glutathione levels were from the Cayman Chemical Company (Ann Arbor, MI, USA).
2.7. Statistical Analysis
Statistical analysis of the present study data was performed by a biostatistician using Shapiro-Wilk, Kruskal-Wallis, Tukey, and/or Mann-Whitney one-way analysis of variance rank sum tests (SPSS statistical software; SPSS Inc., Chicago, IL, USA). Statistical difference for overall malformation between both diabetic groups (vehicle-treated diabetic and LA-treated diabetic) was calculated using a chi-squared test. Results were considered statistically significant when the p value was smaller than 0.05. Data are presented as means SEM.
3.1. Glucose and Body Weight
The average glucose levels in random blood samples from diabetic pregnant rats were significantly higher () (five times) than comparable samples obtained from both groups of control pregnant rats (Table 1).
| higher () than those from both groups of control rats.|
higher () than those from both groups of control and LA-TD rats.
lower () than those from both groups of control and LA-TD rats.
Furthermore, the average glucose levels in samples obtained from LA-treated diabetic (LA-TD) rats were also significantly higher () than samples obtained from both groups of control rats. Daily injections of LA did not normalize blood glucose levels of diabetic rats to the control rats and did not significantly affect the blood sugar concentrations of both diabetic and control rats.
As shown in Table 1, supplementation of control rats with LA (30 mg/kg) for 15 days did not have any effect on body weight gain. However, supplementation of diabetic rats with LA was found to improve body weight gain compared to V-TD group but did not normalize weight gain to both groups of control rats.
3.2. Gross Morphological Study
The mean number of implantation surgeries per litter was similar in all studied groups (Table 1).
In diabetic rats, LA supplementation significantly reduced () resorption rates from in V-TD rats to zero in LA-TD rats.
Fetal body mean weight in V-TD rats was significantly decreased () in comparison to both groups of control and LA-TD rats.
Placental mean weight from both groups of diabetic rats was significantly higher () than those in V-TC rats. In contrast, membrane mean weight in both groups of control rats was significantly higher () than those in both diabetic rat groups.
While no gross morphological malformations were detected in the fetuses of the both groups of control mothers, experimentally induced diabetes was associated with the presence of malformations (Figure 1).
To estimate embryonic development, we measured fetal crown-rump length (CRL, Table 2). In offspring of V-TD rats, a significant decrease () in mean CRL was observed, in comparison with offspring of V- and LA-TC or LA-TD rats. Supplementation of diabetic rats with LA increased CRL in their offspring and was able to normalize the measure to those in V- and LA-TC rats.
| lower () than those from both groups of control and LA-TD rats.|
higher () than those from V-TC rats.
higher () than those from both groups of diabetic rats.
The difference in the mean values of the two groups is statistically significant difference using Shapiro-Wilk, Kruskal-Wallis, and Tukey test one way analysis of variance. The data represent mean SEM.
Administration of LA to diabetic rats for 15 days during gestation marginally reduced the occurrence of malformations (Table 3). The incidence of malformation in V-TD reached approximately 67% (32 of 48 litters) as compared to 42% in LA-TD (25 of 59 litters). Thus, the overall malformation variance between both diabetic groups did not reach statistical significance .
| higher () compared with malformed LA-TD.|
higher () compared with V-TD rats (nonmalformed and malformed).
to increase more than malformed V-TD ().
There are no significant relationships between placental and membrane body weights of any groups.
The difference in the mean values of the two groups is statistically significantly different using Shapiro-Wilk and Mann-Whitney one way analysis of variance rank sum test.
In the V-TD rats, 41.66% of fetuses had growth retardation, as represented by CRL, 39.58% had short mandibles, 31.25% had open neural tubes, 12.5% had malrotations, and 2.08% had short tails. However, in LA-TD rats, the corresponding proportions were 15.25%, 10.17%, 13.56%, 5.09%, and 0.0%, respectively. Some fetuses exhibited a single malformation, while others had multiple anomalies.
3.3. Total Glutathione Levels
To assess the ability of LA to reduce the maternal oxidative stress associated with diabetes, we measured the level of total glutathione in the maternal liver and plasma.
As seen in Table 4, maternal liver of the V-TD rats was associated with significantly reduced () glutathione content in comparison with those in V- and LA-TC rats.
| lower () compared with both groups of control and LA-TD rats.|
increased () compared with those in V-TD rats.
lower () compared with both groups of control rats.
Supplementation of diabetic rats with LA leads to a significant increase () in liver total glutathione in comparison with V-TD rats, although such supplementation failed to normalize glutathione content to those in the liver of both groups of control rats. Furthermore, LA treatment significantly increased () the content of total glutathione in the plasma of diabetic rats compared with those in V-TD rats. This value was not found to be significantly different from that observed in both groups of control rats.
To find out whether maternal diabetes decreases fetal total glutathione levels and if the LA supplementation reverses fetal total glutathione levels, we measured the level of total glutathione in fetuses of all groups (both control and diabetic).
The total glutathione contents in fetuses of V- and LA-TD rats were 51.17% and 45.17% less than in those of V- and LA-TC rats, respectively. On the other hand, the contents of total glutathione in fetuses of both diabetic rats were significantly lower () than those in both control groups.
LA supplementation to diabetic rats significantly increased () total glutathione content of their fetuses as compared with fetuses from V-TD rats.
3.4. Cyclooxygenase Activity in Fetuses
The activity of COX-1 in the fetuses, placentas, and membranes from all control and diabetic mothers represented a small fraction of total COX activity compared with that of COX-2. The presence of a COX-1 inhibitor in the V-TC and V-TD rats was investigated and found to be negative.
The activity of COX-1 showed no significant variations among fetuses from all groups (Table 5). However, the activities of COX-2 in malformed fetuses from V-TD and LA-TD were significantly lower () compared with fetuses from V-TC, LA-TC, nonmalformed V-TD, and nonmalformed LA-TD mothers. Statistically, there was no significant difference between both control and nonmalformed diabetic fetuses.
| lower () compared with fetuses from V-TC, LA-TC, and nonmalformed groups of both groups of diabetic rats.|
3.5. Cyclooxygenase Activity in Placentas
There was a significant decrease in the activities of both COX isoenzymes in the placentas of malformed fetuses from both diabetic pregnancies. A significant decrease () of COX-1 activity was associated with placentas from malformed V-TD and LA-TD fetuses compared with placentas of nonmalformed fetuses (Table 6).
| lower () compared with placentas of nonmalformed fetuses of both groups of diabetic rats.|
higher () compared with placentas of both nonmalformed and malformed fetuses of diabetic rats.
higher () than those in malformed fetuses from V-TD rats.
A similar trend was found in activities of COX-2; the activities of COX-2 in V-TD and LA-TD placentas from malformed fetuses were significantly lower () than nonmalformed fetuses.
The activities of COX-2 in the placentas from both control fetuses were significantly higher () compared with those from both nonmalformed and malformed fetuses of diabetic mothers.
Supplementation of diabetic rats with LA leads to a significant increase () in the activities of COX-2 in the placentas from malformed fetuses in comparison with those from V-TD rats. Thus, LA treatment to diabetic mothers failed to normalize activities of COX-2 to the control rats.
3.6. Cyclooxygenase Activity in Membranes
The total activity of COX in membranes from the both diabetic was significantly lower () in comparison to membranes from both the control fetuses (Table 7). Furthermore, the activity of total COX in membranes from nonmalformed fetuses of LA-TD was significantly higher () compared with the nonmalformed fetuses of V-TD mothers. The activities of COX-1 in membranes showed no significant variations among fetuses from all groups.
| lower () in comparison to membranes from both control fetuses.|
lower () than those in the both control fetuses.
higher () compared with the nonmalformed fetuses of V-TD rats.
lower () than those in the membranes of fetuses from V-TC, LA-TC, and nonmalformed fetuses from both groups of diabetic rats.
higher () than those in nonmalformed fetuses from V-TD rats.
The activities of COX-2 in membranes of malformed fetuses from V-TD and LA-TD were significantly lower () than those in the membranes of fetuses from V-TC, LA-TC, and nonmalformed fetuses from V-TC and LA-TD mothers. Furthermore, the activity of COX-2 in membranes from nonmalformed fetuses of LA-TD was significantly higher () than those in nonmalformed fetuses from V-TD rats.
3.7. Levels of PGEM and PGFM in Fetuses
Decreased levels of PGEM and elevated levels of PGFM in the fetuses, placentas, and membranes were characteristic of both experimental diabetic gestation groups associated with malformation.
As shown in Table 8, the mean levels of PGEM in fetuses (nonmalformed and malformed) from both groups of diabetic mothers were significantly lower () than those in both control groups.