Maternal use of flaxseed oil during pregnancy and lactation prevents morphological alterations in pancreas of female offspring from rat dams with experimental diabetes

Abstract

Nutritional recommendations have promoted the increased need to consume n-3 polyunsaturated fatty acids. Flaxseed is the richest dietary source of n-3 fatty acids among plant sources and is widely used for its edible oil. This study aimed to investigate whether maternal use of flaxseed oil has effects on pancreas morphology in the female offspring of diabetic mothers. Female Wistar rats (n = 12) were induced into diabetes by a high-fat diet and low dose of streptozotocin. After confirmation of the diabetes, rats were mated, and once pregnancy was confirmed, they were allocated into three groups (n = 6): high-fat group (HG); flaxseed oil group (FOG); and control group (CG) (non-diabetic rats). At weaning, female offspring (n = 6/group) received standard chow diet. The animals were euthanized at 180 days. Pancreas was collected for histomorphometric and immunohistochemistry analysis. HG showed hypertrophy of pancreatic islets (P < 0.0001), whereas FOG offspring had islets with smaller diameters compared to HG (P < 0.0001). HG offspring showed higher percentage of larger (P = 0.0061) and lower percentage of smaller islets (P = 0.0036). HG showed lower islet insulin immunodensity at 180 days (P < 0.0001), whereas FOG was similar to CG (P < 0.0001). Flaxseed oil reduced the damage caused by maternal hyperglycaemia, promoting normal pancreas histomorphometry and β-cell mass in female offspring.

Diabetes mellitus (DM) is a group of metabolic disorders characterized by chronic hyperglycaemia caused by lack of insulin production by pancreatic beta-cells or by defects in insulin receptors on target cells, resulting in hyperglycaemic metabolic disease (American Diabetes Association 2014). DM is a chronic disease present throughout the world, currently affecting approximately 382 million people. The number of people with the disease is set to rise beyond 592 million in <25 years (International Diabetes Federation 2013). As the incidence of diabetes continues to increase and affect individuals of all ages, including young children, women of childbearing age are at increased risk to develop this disease during pregnancy (Hunt & Schuller 2007; Nolan et al. 2011).

Pregnancy is a diabetogenic situation per se resulting from metabolic adaptations performed by the maternal metabolism to ensure adequate food support to the foetus (physiological insulin resistance). In predisposed women, this can lead to gestational diabetes. Women with gestational diabetes, as well as women with pregestational diabetes (type 1 or type 2), are classified as high-risk pregnancies, and their children show increased morbidity and mortality in the perinatal period (Plagemann 2011).

Several studies have observed that insults occurring during intrauterine life are associated with several abnormalities, both functional and structural, at adulthood. Diabetes is a complication that, when it occurs during pregnancy, can substantially influence the development of offspring during foetal and postnatal life (Poston 2010; Plagemann 2011; Yessoufou & Moutairou 2011).

Diabetes during pregnancy induces changes and adaptation in foetal pancreatic activity, in response to the increased supply of glucose from mother to foetus. Chronic hyperglycaemia in the intrauterine environment in pregnant rats induces hyperglycaemia and hypoinsulinaemia in the offspring at birth and is associated with foetal growth retardation. The foetal pancreas weight is decreased whereas the percentage of endocrine tissue is increased (Holemans et al. 2003; Fetita et al. 2006), and young adults exhibit mild glucose intolerance that worsened with time (Fetita et al. 2006).

It is imperative to mention that the foetal programming of chronic diseases has important medical and economic ramifications. Recently it has been proposed that interventions during pregnancy and lactation may be effective in preventing disease in adulthood (Gluckman & Hanson 2004).

With regard to therapy, researchers have suggested that n-3 polyunsaturated fatty acids (n-3 PUFAs) have the ability to prevent disease. Several benefits of n-3 PUFAs in relation to diabetes have been implicated, such as improvement in glucose homoeostasis, which involves several molecular mechanisms, such as stimulation of insulin signalling pathway (Wu et al. 2012). Polyunsaturated fatty acids of the n-3 family include eicosapentaenoic acid (EPA – 20:5n-3) and docosahexaenoic acid (DHA – 22:6n-3) from seafood and alpha-linolenic acid (ALA, 18:3n-3) from plant sources (Wu et al. 2012). Flaxseed has 57% of n-3, 16% of n-6, 18% of monounsaturated fatty acids and only 9% of saturated fatty acids (Morris 2007; de Almeida et al. 2009). Flaxseed is the seed of the flax plant (Linum usitatissimum L.) which is a member of the Linaceae family (de Almeida et al. 2009; Shim et al. 2014) and is widely used for its edible oil in many parts of the world (Naqshbandi et al. 2013).

Due to the wide dissemination in the general media of the relationship between food and health, societal concern with functional food has increased exponentially. Functional foods are any food or ingredients that, beyond basic nutritional functions, when consumed as part of the usual diet, produce metabolic and/or physiological effects and/or health benefits, and should be safe for consumption without medical supervision (Guarda et al. 2014). Therefore, flaxseed, besides its by-products, such as its oil, can be framed as a functional food because it contains physiologically active components that may promote health benefits beyond basic nutrition.

This study was designed to test whether maternal flaxseed oil use during pregnancy and lactation has an effect on glucose metabolism and pancreatic morphology of adult female offspring of diabetic Wistar rat.

Materials and methods

Female Wistar rats (3 months old) from the Centre of Laboratory Animals of Fluminense Federal University were housed under a controlled temperature (21 ± 1°C), humidity (60 ± 10%) and 12-h light/dark cycle, with free access to water and food.

Experimental diabetes induction

Non-diabetic female Wistar rats (n = 12) were fed with a high-fat diet (60% of energy from lipid, 14% from protein and 26% from carbohydrates) for an initial period of three weeks. The other six rats were fed a control diet based on casein (10% of energy from lipid, 14% from protein and 76% from carbohydrates), both ad libitum, thus forming two groups: high-fat group (HG) (n = 12) and control group (CG) (n = 6). After three weeks of the high-fat diet, the rats of HG received intraperitoneal injection of streptozotocin (STZ) (Sigma Chemical, St. Louis, MO, USA) at a lower dose (35 mg/kg) dissolved in vehicle (sodium citrate buffer 0.01M, pH = 4.5) (Srinivasan et al. 2005; Correia-Santos et al. 2012). The rats that consumed the control diet received the vehicle solution intraperitoneally only. The groups continued to receive their respective experimental diets for another week, totalizing a sum of four weeks of exposure to the dietary standards adopted. At the end of the fourth week, the blood was collected to confirm the diabetes. The confirmation was established through the plasma glucose concentration, which was above 300 mg/dl (Srinivasan et al. 2005).

Experimental design

All rats were mated overnight with non-diabetic male Wistar rats at a ratio of two females per male. The mornings on which spermatozoa were detected on vaginal swabs were established as day zero of pregnancy. The mating procedure continued for 15 consecutive days, which comprises approximately three oestrous cycles.

After confirming the pregnancy, the rats were placed in individual cages and allocated into three experimental groups: high-fat group (HG) (n = 6), of diabetic pregnant rats, which received a high-fat diet [49% of energy (Table1)]; flaxseed oil group (FOG) (n = 6), of diabetic pregnant rats, which received a high-fat diet, in which soya bean oil was replaced by flaxseed oil; and control group (CG) (n = 6), of non-diabetic rats, which received a control diet based on casein. The dietary standard cited was adopted throughout the period of gestation and lactation. During pregnancy and lactation, the macronutrient values of the diets were modified to include 19% of energy from protein to meet nutritional needs during this period. The content of vitamins and minerals followed the recommendations of the AIN-93G (Reeves et al. 1993). The experimental diets were prepared at the Laboratory of Experimental Nutrition from Fluminense Federal University. The composition can be seen in Table1. After delivery, the offspring was fed with a normal lipid and protein diet (Nuvilab®, Nuvital LTDA, Paraná, Brazil) containing 22% protein (main protein sources are meat, fish, soya and amino acids), 66% carbohydrate and 11% lipid until they were 180 days old.

Table 1.

Nutritional composition of experimental diets during pregnancy and lactation

Nutrients (g/kg)	Diets
	Control	High-fat	High-fat with flaxseed oil
Casein (≥85% of protein)*	190	230	230
Corn starch*	539,486	299,486	299,486
Sucrose‡	100	100	100
Soybean oil§	70	70	0
Flaxseed oil ‡‡	0	0	70
Lard‖	0	200	200
Fibre (cellulose)¶	50	50	50
Vitamin mix (AIN-93G)**	10	10	10
Mineral mix (AIN-93G)**	35	35	35
Cystine††	3	3	3
Choline††	2.5	2.5	2.5
Terc butyl hydroquinone	0.014	0.014	0.014
Total	1000	1000	1000
Carbohydrate (% of total kcal)	64	32	32
Protein (% of total kcal)	19	19	19
Fat (% of total kcal)	17	49	49
Energy (Kcal/kg)	3950	4950	4950

AIN-93G, American Institute of Nutrition-93G. Ingredients used in diets preparation were provided by

^*Comércio e Indústria Farmos Ltda. (Rio de Janeiro RJ, Brazil).

^‡União (Rio de Janeiro, RJ, Brazil).

^§Liza da Cargill Agricultura Ltda. (Mairinque, SP, Brazil).

^¶Microcel da Blanver Ltda. (Cotia, SP, Brazil).

^‖Sadia Comercial Ltda.

^**PragSoluções Comércio e Serviços Ltda-ME (Jaú, SP, São Paulo).

^††M. Cassab Comércio e Indústria Ltda. (São Paulo, SP, Brazil).

^‡‡Giroil Agroindustria LTDA (Santo Ângelo, RS, Brazil).

After weaning, the pups (18 females) were kept in individual cages until they reached 180 days of life. Food intake and body mass of groups were measured once a week. The energy intake per gram of body mass gained, termed the feed efficiency (FE), was calculated as a digestive and metabolic indicator of the ease that energy consumed was added as body mass. Food intake (g/day) was multiplied by the dietary energy (kcal/g) to obtain the daily energy intake per rat. Feed efficiency was calculated as [(body mass gain/kcal intake) × 100] (Barbosa-da-Silva et al. 2012).

Metabolic analyses

The oral glucose tolerance test (OGTT) was taken 1 week before the euthanasia. For blood glucose levels, blood was obtained by milking the tail after a small incision had been made. (glucometer Accu-chek®, Roche Diagnostic, São Paulo, Brazil). The OGTT was performed using a solution of glucose in sterile saline (0.9% NaCl) at a dose of 2.0 g/kg. The glucose was administered by orogastric gavage after a six-hour fasting period. The blood glucose concentration was measured before glucose administration (0 min) and 15, 30, 60 and 120 min after administration. An intraperitoneal insulin tolerance test (IpITT) was also conducted at the same time of OGTT. For the IpITT, fasting blood glucose levels were measured (after a 4-h fasting period using the blood obtained from the tail vein) using a glucometer (Accu-Chek®; 0 min). Insulin was subsequently injected intraperitoneally (1.0 U/kg, Humalog®, insulin lispro, Eli Lilly and Company, Indianapolis, IN, USA), and blood glucose was measured again at 15, 30, 60 and 120 min. The analysis considered the area under the curve (in arbitrary units, a.u.) to evaluate the glucose intolerance and the insulin resistance (prism version 5.03 for windows; GraphPad Software, San Diego, CA, USA).

Blood collection, glucose and insulin levels determination

At the end of 180 days, and after a 6-h fasting period, the offspring of the diabetic and the non-diabetic mothers were anesthetized with an intraperitoneal injection of Thiopentax® (Sodium thiopental 1 g, Cristália Produtos Químicos Farmacêuticos LTDA, Brazil) at 5% (0.15 ml/100 g body weight), and the blood was taken via cardiac puncture and put into tubes without anticoagulant. Blood samples were centrifuged (Sigma centrifuge) at 3.500 g for 15 min to separate the serum. Serum insulin was analysed using Multiplex Biomarker Immunoassays for Luminex xMAP technology (Milipore, Billerica, MA, USA, cat. no. RADPK-81K). Glucose level was determined in whole blood using glucometer and strips for glucose (Accu-Chek; Roche).

Histomorphometry of the pancreas

At the time of euthanasia the pancreas was removed and weighed on an analytical digital balance for comparison of their relative weight, which was calculated by adjusting for 100 grams of body weight. Fragments of the pancreas (6/group) were rapidly fixed (freshly prepared fixatives, 1.27 mol/l formaldehyde in 0.1 M phosphate buffer, pH 7.2) for 48 h at room temperature. The material was embedded in paraffin, sectioned into 5-μm-thick slices and stained with haematoxylin and eosin. The slices were scanned with the help of the capture slices device, Aperio ScanScope (ScanScope CS2, Leica Biosystems Imaging, Inc., Vista, CA, USA), and the pancreatic islets were quantified. The largest and smallest diameters of each islet were measured with a ruler to calculate the average diameter of islets (Image Scope, version 11.2.0.780, Aperio Technologies, Leica Biosystems Imaging, Inc., Vista, CA, USA). At least 50 islets per rat were analysed (Gundersen & Jensen 1985). The islets were separated by diameter into small (<125 μm) and large (>150 μm) (MacGregor et al. 2006). The pancreatic islet (PI) density, expressed as PI/mm², was calculated as a ratio of the PI number divided by the area of parenchyma evaluated in 20 microscope fields (Saldanha et al. 2001). The islet mass was obtained by multiplying the pancreatic islet density by the pancreatic mass (Fernandes-Santos et al. 2013).

Immunohistochemistry

Sections obtained through the entire pancreas (six sections/group) were deparaffined, and antigen retrieval was formed with citrate buffer (pH 6.0) for 45 min at 96° C. Then, endogenous peroxidase blockade was prepared with hydrogen peroxidase (3% H₂O₂ in H₂O) and finally was inhibited with phosphate-buffered saline/bovine serum albumin (5%). Sections were incubated with anti-insulin antibody (dilution 1:500, CMC27311021, Cell Marque, Rocklin, CA, USA) and were treated with a biotinylated secondary antibody (DCMT-999, REVEAL Complement, Spring Bioscience, California, USA); the immunoreactions were amplified with biotin–streptavidin system (DHRR-999, REVEAL HRP Conjugate, Spring Bioscience, California, USA). Immunostaining was visualized after incubating the section with 3,3 diaminobenzidine tetra-chloride (DakoCyntomation, Glostrup, Denmark) and counterstaining with Mayer haematoxylin.

Digital images from pancreatic islets (10 pictures/animal/group) were obtained and studied. The beta-cell immunodensity was estimated via image analysis using density threshold selection tool applied to the insulin-positive areas of the islets expressed as percentage of the islet (ImageJ 1.47v, Wayne Rasband, National Institutes of Health, USA). Thus, beta-cell mass was estimated as the product of beta-cell immunodensity and the islet mass (Fernandes-Santos et al. 2013).

Statistical analysis

Data were reported as the mean ± standard error of the mean. Data were tested for normality and homogeneity of variances, and the differences among groups were tested, when appropriated, with one-way analysis of variance (anova), followed by a Tukey's post hoc test. P value ≤0.05 was considered statistically significant (graphpad prism v. 5.04 for windows, GraphPad Software).

Results

Ethical Approval Statement

Animal protocol was approved by the Animal Ethics Committee of the Fluminense Federal University (Protocol Number CEA 035/2010), and the procedures were in accordance with the guidelines for experimentation with animals (NIH Publication N°. 85-23, revised 1996).

Dams

During pregnancy and lactation, the levels of maternal glucose of the studied groups were CG: 90 ± 5.4 mg/dl, HG: 438.5 ± 3.5 mg/dl and FOG: 405.7 ± 14.8 mg/dl. We noticed that diabetic rats had higher levels of glucose than control rats (P < 0.0001).

Offspring

Body mass, food intake and feed efficiency

Maternal hyperglycaemia led to low birthweight in the offspring of diabetic mothers. At birth, CG, HG and FOG weighed 6.8 ± 0.4 g, 5.4 ± 0.4 (−20.6%) and 5.3 ± 0.2 g (−22.1%) (P = 0.0107) respectively. At weaning, the HG and FOG offspring remained lighter than CG (−33.5% and −49.1%, respectively, P < 0.0001) and when we compared FOG to HG, it was 23.5% lighter (P < 0.0001, Table2). As shown in Figure1, HG and FOG offspring remained with lower weight than CG until the 77th day; from this time onwards, the growth curve was similar between groups and was not found to be different between the three groups with respect to body weight at 180 days (Table2).

Table 2.

Biometric data, feed intake and feed efficiency of the experimental groups throughout the experiment

	CG	HG	FOG	P value
Weaning weight – 21 days (g)	44.8 ± 1.4	29.8 ± 1.1[a]	22.8 ± 2.7(a,b)	<0.0001
Final weight – 180 days (g)	276.8 ± 7.4	255.3 ± 5.4	262.9 ± 12.3	0.2514
Weight variation (g)	231.9 ± 6.1	219.8 ± 5.0	240.1 ± 10.3	0.19
Cumulative feed intake (g)	3053 ± 97.6	3481 ± 242.8	3169 ± 145.6	0.2271
Feed intake (g/day/animal)	19.2 ± 0.6	21.9 ± 1.5	19.9 ± 0.9	0.2122
Feed efficiency (%)	2.1 ± 0.1	1.8 ± 0.1	2.1 ± 0.1	0.0802

The data were expressed as mean ± SEM. Groups: CG, control group; HG, high-fat group; and FOG, flaxseed oil group. The letter (a) represents statistical difference compared to the control group and (b) statiscal difference compared to the high-fat group (one-way anova, P < 0.05).

Figure 1.

Body weight of female offspring during the 180 days of experiment. Groups: CG, control group, HG, high-fat group and FOG, flaxseed oil group. The star (*) represents statistical difference among groups HG and FOG compared to CG (One-way anova, P < 0.05).

We observed that maternal hyperglycaemia did not affect feed intake during the study period: the rats of the three experimental groups consumed the same amount of food and daily consumption, and feed efficiency were also not affected, being similar in groups (Table2).

Evaluation of glucose metabolism

Fasting glucose and insulin levels were similar in the three groups at 180 days (P = 0.3795 and P = 0.2460, respectively, Table3). To investigate the ability of the groups to regulate glucose metabolism, OGTT and IpITT were performed. At 180 days, the three groups did not exhibit glucose intolerance, showing normal responses to glucose infusion (OGTT) (P = 0.5268, Table3), and normal degrees of insulin resistance, because the groups showed no differences in the IpITT area under the curve (P = 0.1272, Table3).

Table 3.

Metabolic and pancreatic parameters of experimental groups at 180 days

	CG	HG	FOG	P value
Glucose (mg/dl)	106.8 ± 2.7	101.6 ± 2.3	105.6 ± 2.3	0.3795
Insulin (μIU/ml)	38.6 ± 9.6	25.0 ± 5.1	43.1 ± 5.5	0.2460
OGTT (AUC, u.a.)	14490 ± 376.0	14653 ± 272.9	14049 ± 495.0	0.5268
IpITT (AUC, u.a.)	6673 ± 183.4	7265 ± 807.2	5973 ± 351.8	0.1272
Pancreas absolute mass (g)	2.68 ± 0.23	2.03 ± 0.10(a)	2.69 ± 0.28(b)	0.0447
Pancreas relative mass (%)	0.964 ± 0.07	0.801 ± 0.04(a)	1.012 ± 0.07(b)	0.0469
Islet diameter (μm)	90.4 ± 2.0	110.2 ± 2.2(a)	88.1 ± 2.3(b)	<0.0001
Islet numeric density (PI/mm²)	1.76 ± 0.13	1.99 ± 0.19	1.64 ± 0.21	0.4156
Islet mass (mg)	472.6 ± 35.6	403.1 ± 38.0	439.9 ± 57.9	0.5369
Islet insulin immunodensity (%)	49.64 ± 0.75	39.79 ± 1.46(a)	52.69 ± 1.45(b)	<0.0001
Beta cell mass (mg)	243.6 ± 17.7	160.4 ± 15.2(a)	231.8 ± 30.5(b)	0.0325

a.u., arbitrary units; AUC, area under the curve; OGTT, oral glucose tolerance test; IpITT, intraperitoneal insulin tolerance test; PI, pancreatic islet. The data were expressed as mean ± SEM. Groups: CG, control group; HG, high-fat group; and FOG, flaxseed oil group. The letter (a) represents statistical difference compared to the control group and (b) statiscal difference compared to the high-fat group (One-way anova, P < 0.05).

Pancreas weight and pancreatic histomorphometry

Maternal hyperglycaemia led to a lower absolute and relative pancreas weight in HG (−24.2%, P = 0.0447, −16.9%, P = 0.0469, respectively, Table3). However, maternal use of flaxseed oil led to pancreas absolute and relative weight similar to the CG (Table3). With regard to the morphology, offspring from HG had bigger pancreatic islet diameter when compared to the CG (+21.9%, P < 0.0001, Table3). FOG offspring showed smaller islet diameter when compared to the HG (−20%, P < 0.0001, Table3). When the islets were separated by diameter into small and large, the HG showed lower percentage of small islets when compared to the CG (CG: 79.8 ± 1.8%; HG: 66.1 ± 2.9%, −17.2%, P = 0.0036, Figure2) and FOG had a higher percentage when compared to the HG (FOG: 81.1 ± 4.1%, +22.7%, P = 0.0036, Figure2) and similar to the CG. Regarding the islets considered large in diameter, we noted that the HG offspring had a higher percentage when compared to the CG (CG: 13.8 ± 2.1%, HG: 24.0 ± 2.4%, +73.9%, P = 0.0061, Figure2) and once again, the percentage of large islets of FOG was similar to the CG and lower than HG (FOG: 13.0 ± 2.4%, −45.8%, P = 0.0061, Figure1). Analysing the pancreatic islets density we found no differences between groups at 180 days (P = 0.4156, Table3), as well as the islet mass, which no difference was found between the groups (P = 0.5369, Table3).

Figure 2.

Size distribution of pancreatic islets at 180 days. CG, control group, HG, high-fat group and FOG, flaxseed oil group. The letter (a) represents statistical difference compared to the control group and (b) statiscal difference compared to the high-fat group (One-way anova, P < 0.05).

Insulin immunodensity

There were positive immunoreactions for insulin (β-cells) in pancreatic islets in all the groups. The characteristic positive immune reaction of insulin was predominantly in the central region of the islet (Figure3). The islets of HG offspring showed decreased insulin immunodensity when compared to the CG (−19.8%, P < 0.0001), whereas FOG insulin immunodensity was similar to the CG and higher than HG (+32.4%, P < 0.0001, Table3). Consequently, beta-cell mass was decreased in HG offspring in comparison with CG (−34.1%, P < 0.0325, Table3) while FOG was similar as CG (Table3).

Figure 3.

Microscopic images of the pancreas tissue immunostained with anti-insulin antibody (beta cells) in 6-month-old (180 days) female offspring. (a), Control group; (b), High-fat group; (c), Flaxseed oil group (20×).

Discussion

This work shows that female offspring from mothers with severe hyperglycaemia during pregnancy and lactation have low birthweight and low weight at weaning, besides changes in pancreatic structure when adults. For the first time, we show that maternal use of flaxseed oil had beneficial effects in the female offspring from diabetic mothers such as avoiding hypertrophy and improving pancreatic islet β-cell expression.

A very common injury seen in cases of severe maternal hyperglycaemia is low birthweight (Holemans et al. 2003; Fetita et al. 2006; Song et al. 2012), which agrees with our findings, where the female offspring from HG diabetic mothers were lighter than CG, as well as FOG. The low birthweight can be explained by the fact that during the pregnancy of the diabetic mothers, the foetus is confronted with severe intrauterine hyperglycaemia, which induces foetal islet hypertrophy and β-cell hyperactivity, a phenomenon that may result in early hyperinsulinaemia. This overstimulation of foetal β-cells limits their adaptation, so they become depleted of insulin granules and incapable of secreting insulin. β-Cell exhaustion results in foetal hypoinsulinaemia. Hypoinsulinaemia and a reduced number of insulin receptors in target cells lead to a reduction in the foetal glucose uptake. The growth of foetal protein mass is suppressed, and the foetal protein synthesis is consistently lowered, leading to a foetal microsomia. Postnatal development is retarded, and these offspring remain small in adulthood (Holemans et al. 2003; Yessoufou & Moutairou 2011).

Epidemiological and experimental studies have reported that fish oil supplementation, a source of n-3 PUFA, when given during pregnancy, can increase birthweight and thus reduce the chances of developing chronic diseases in adulthood (McGregor et al. 2001; Olafsdottir et al. 2005). Some mechanisms may explain this benefit; among them, it is postulated that the vasodilator power of DHA increases the intrauterine placental flow (Rogers et al. 2004), and therefore the avidity in nutrients and oxygen supply to the foetus, which elicit the increase of birthweight. The efficiency of conversion of ALA to its long-chain derivatives remains controversial and requires further extensive scientific research. Some studies in humans using stable isotopes suggest that most of the ALA from the diet is readily β-oxidized and used as energy substrate, being limited in its enzymatic conversion to EPA (0.2 a 8%) and DHA (<0.05 a 4%) (Burdge 2006; Plourde & Cunnane 2007). Conversely, in contrast to fish oil, which has DHA already formed in its composition, the n-3 from flaxseed oil must be converted to EPA and DHA, and due to this, weight gain that was expected in the FOG did not occur because the supply of DHA was lower in this group.

At weaning, day 21, all female offspring from diabetic mothers were still lighter than CG. Guarda et al. (2014) offered high-fat diet with flaxseed oil to healthy Wistar rats during lactation and also observed low weight of male and female offspring at weaning compared to offspring of mothers who consumed control diet. The administration of high-fat diet with 19% of flaxseed oil during lactation changed the milk composition, with lower cholesterol and triacylglycerol content, and hence, they concluded that low weight at weaning was due to this factor. As our animals also received high-fat diet with flaxseed oil during lactation, we conclude that this modification also occurred in milk. Another reason is that streptozotocin used to induce diabetes in rats leads to a reduced ability of the mammary gland to synthesize fatty acids, leading to lower amount in milk (Jackson et al. 1994; Blondeau et al. 2011). Because of this, changes in milk composition may have contributed to the growth retardation observed after birth in offspring of mothers with hyperglycaemia.

A fairly frequent phenomenon observed in cases of intrauterine growth restriction is accelerated postnatal growth (catch-up) to compensate the low birthweight. Accordingly, the animals become more susceptible to increased risk of developing type 2 diabetes and metabolic syndrome in adulthood (Hales & Ozanne 2003). The females of the two groups of diabetic mothers, who were lighter at weaning than the CG, succeeded to match their weights to CG by their 70th day, recovering its growth curve, leaving it similar to CG, even consuming the same amount of food, indicating a possible catch-up postweaning.

Regarding food intake, we observed that the addition of flaxseed oil to the high-fat diet did not affect the food intake of the offspring throughout life. For these offspring, increase in food intake, which causes obesity and insulin resistance was expected in the long run. Foetal hyperinsulinaemia contributes to the dysfunction of key critical/essential pathways for normal development of the hypothalamic neural networks for energy balance (Plagemann 2011). Although no significant differences were observed, female HG offspring consumed 14% more food than the CG, indicating a possible increase in the expression of orexigenic peptides and decrease in the expression of anorexigenic peptides as a result of the changes caused by foetal hyperinsulinaemia.

Animal models have shown convincingly that diabetes can be transmitted by intrauterine exposure to maternal hyperglycaemia. Maternal hyperglycaemia during critical periods of development has been linked to reduce insulin secretion in response to the administration of glucose (Fetita et al. 2006), However, in this study, severe maternal hyperglycaemia did not affect glucose tolerance in female offspring when measured at 180 days after birth. The same result was found by Zhao & Weiler 2010), where maternal hyperglycaemia did not affect the glucose tolerance of the offspring of Sprague Dawley rats, in both genders, at three months of age. Similar to our study, Song et al. (2012) observed that when kept on standard chow diet, offspring from diabetic mothers exhibited relatively normal glucose tolerance, resembling their counterparts from mothers with normal glucose. As with OGTT, we found no differences between the groups with regard to IpITT, a method used to measure peripheral insulin resistance, at 180 days. Blondeau et al. (2011) evaluated the glucose metabolism of the offspring of Sprague Dawley with diabetes during pregnancy and lactation at 3, 6 and 12 months and found insulin resistance through IpITT test only at 12 months of life of these animals. Similar to our results, at 6 months no differences were found in the IpITT area under the curve between the groups, but we realized that HG had an area under the curve 8.9% bigger than CG. Maybe if the study were taken longer, we would find insulin resistance in these rats.

Concerning fasting glucose, Zeng et al. (2010) also found no differences in fasting glucose levels between the CG and the offspring from Wistar rats with severe hyperglycaemia at six months of age. These observations agreed with Blondeau et al. (2011), who observed that fasting glucose, and insulin levels were similar among male offspring of diabetic rats and healthy rats at three and six months of age. Another study, evaluating the effects of severe hyperglycaemia upon fasting glucose and insulin levels in male rats at six months, observed no differences between the groups originating from diabetic mothers and control mothers (Song et al. 2012). We emphasize that in most studies, only male offspring have been analysed, making it difficult to compare our results with the male offspring of diabetic rats.

It has been described in the literature that hyperplasia of pancreatic islets can occur due to maternal hyperglycaemia (Holemans et al. 2003; Fetita et al. 2006) through a possible neogenesis during the perinatal period, which can be observed in adulthood. Analysing the pancreatic islets density in the experimental groups we noted that the groups did not differ among themselves; however, maternal hyperglycaemia increased the number of pancreatic islets in HG, inasmuch as that the females of this group had 13.1% more pancreatic islets than the CG. It is noteworthy that the use of flaxseed oil did not lead to an increased number of islets, as the females offspring from this group have −17.6% of islets compared to HG.

According to Remacle et al. (2007), the offspring of diabetic mothers have hypertrophy of pancreatic islets due to the hyperglycaemic intrauterine environment, resulting from overstimulation of these islets. Analysing the average diameter of the pancreatic islet, we observed that this was affected by maternal hyperglycaemia, where HG has a larger diameter than CG. In contrast, Song et al. (2012), studying male offspring from Sprague Dawley rats with severe hyperglycaemia during pregnancy and lactation, observed no differences in the size of the pancreatic islet between groups derived from diabetic mothers fed a control chow after weaning and control mothers at six months of age. We emphasize a greater protective effect of flaxseed oil on the ability to prevent pancreatic islet hypertrophy, because the diameters of the islets were smaller than those of HG and were similar to the CG at 180 days. It is well known that n-3 PUFAs activate the peroxisome proliferator-activated receptors (PPAR) and the expression of isoform PPAR γ in β-cells controls the expression of genes involved in glucose metabolism. Hence we expect that n-3 reduces basal overstimulation of pancreatic β-cells, which occur in the offspring of mothers with diabetes during pregnancy, from birth to adulthood (Plagemann 2011), not leading to hypertrophy of the islets.

It has been reported in the literature that when pancreatic islets are isolated from healthy rats the percentage of small islets is higher than the percentage of large islets (MacGregor et al. 2006). This is similar to the results observed in our study, where all groups of females offspring had a higher percentage of small islets. When comparing between groups, it was observed that HG had a higher percentage of large islets and a lower percentage of small islets than CG at 180 days. The largest amount of large islets in HG is due to the overstimulation that these islets had at the time of pregnancy, where it was confronted with a severe maternal hyperglycaemia in consequence of the need to produce more insulin. This condition led to increase in their cells and thereby increasing the size of islet, and this characteristic was maintained until adulthood (Fetita et al. 2006; Remacle et al. 2007). Once again, we emphasize the effect of the use of flaxseed oil, because its use has not led to this situation, as the FOG had the distribution relative to the size of pancreatic islet similar to CG.

Although HG animals showed enhanced numerical density of pancreatic islets and islet size, the absolute and relative weight of the pancreas was lower than in the other groups. Holemans et al. (2003) reported that the foetal pancreas weight is decreased in offspring of diabetic mothers, although the percentage of endocrine tissue is increased, endorsing our findings and indicating a lower percentage of exocrine tissue in detriment to the amount of endocrine tissue. Similar to CG, the FOG showed the same absolute and relative weight of the pancreas, and all parameters relative to the endocrine part were equivalent to those seen in CG.

Large pancreatic islets, as found in HG offspring, secrete less insulin, and an explanation could be that these have lower beta-cells immunodensity per islet and less insulin per cell (Fujita et al. 2011; Huang et al. 2011). Our results of insulin immunodensity corroborate this idea, because the HG offspring showed lower density of immunostaining when compared to other groups, as well as decreased beta-cell mass. Once more, we observed the effect of flaxseed oil in avoiding this situation, because the density of the immunostaining and the beta-cell mass was greater than HG and similar to CG. N-3 LCPUFAs and their metabolites are natural ligands of PPAR γ (Edwards & O'Flaherty 2008; Calder 2012), and studies have shown that they have direct beneficial effects on pancreatic β-cells, such as improvements on insulin secretory capacity in pancreatic islets isolated from Wistar rats and hamsters β-cells strains (Van Herpen & Schrauwen-Hinderling 2008). Thus we credit the greater expression of insulin in FOG compared to HG to be linked to this relationship between LCPUFA and PPAR γ.

Conclusion

This study showed that the adverse pancreatic remodelling, characteristic of rats from mothers with severe hyperglycaemia, was prevented by maternal use of flaxseed oil during pregnancy and lactation; and that, more specifically, the effect was associated with enhanced expression of insulin and reduced pancreatic islet hypertrophy in the female offspring.

Conflict of interest

No author has a conflict to report.

Authorship

AMCS, KCLA and GTB designed the study; AMCS, AS, GCV, JSA and ADP generated, collected and assembled the data; AMCS, KCLA and GTB analysed and interpreted the data; AMCS, AS and GCV drafted the manuscript; KCLA and GTB revised the manuscript; AMCS, KCLA and GTB approved the final version of the manuscript.