Nutritional implications of trans fatty acids during perinatal period, in French pregnant women

TFA Nevertheless, for this French population, there was no relation between either weight or head circumference of newborn and the TFA levels in both adipose tissue and plasma lipids of their mother.

Summary : Some studies have demonstrated the transfer of trans fatty acids (TFA) across the human placenta. It was suggested that TFA might disturb the metabolism of essential fatty acids (EFA) in fetus and consequently might affect intrauterine human growth more or less according to the TFA intake level of the mother. In this context, the objective of this study was to assess, for French pregnant women, possible impact of their TFA intake on parameters of their term infants: 1/ TFA composition of the umbilical cord (plasma lipids and parietal phospholipids), 2/ birth weight and head circumference. The TFA composition of maternal and umbilical plasma lipids, and parietal phospholipids of umbilical cord were determined by gas chromatography, associated with a thinlayer chromatography for the phospholipids. Because TFA content of adipose tissue is a reliable biochemical indicator of the usual TFA intake level, maternal adipose tissue was also analyzed. Trans fatty acid (TFA) percentage was significantly higher (p = 0.001) in maternal (0.9% of total fatty acids) than in umbilical plasma total lipids (0.6%) thus confirming their placental passage. Moreover, trans isomer pattern of cord plasma lipids was different from the maternal's one, especially regarding trans isomers of linoleic acid (9c,12c- 18:2). Percentage values of 9c,13t + 9t,12t and 9t,12c isomers were respectively, 2 and 3 times higher (p < 0.001) in umbilical than in maternal plasma (Except trans 16:1 acids). All trans isomers observed in umbilical plasma lipids were detected in both parietal and vessel (vein and arteries) phospholipids of umbilical cord. Nevertheless, 9c13t + 9t12t isomer mix level was significantly higher (p < 0.001) in artery than in vein TPL. Moreover, in artery TPL, percentage value of 9c13t+9t12t isomer mix was inversely correlated (r = -0.703, p = 0.003) with arachidonic acid content. Nevertheless, for this French population, there was no relation between either weight or head circumference of newborn and the TFA levels in both adipose tissue and plasma lipids of their mother.

ARTICLE
Occurrence of trans fatty acids (TFA) in human tissues depends on their availability in the diet, because humans do not synthesize these fatty acid isomers. It is now well known that trans fatty acids are provided by three different dietary sources: the ruminant fats (milk and dairy products, beef, mutton, tallow) [1], the partially hydrogenated vegetable oils (PHVO) (some margarines, shortenings, bakeries and chips) [2] and to a lesser extent, the refined vegetable oils [3].
Several animal studies, in vivo [4][5][6] and in vitro experiments in animal tissues [7][8][9][10] as well as human fibroblast studies [11] have demonstrated that TFAs impair the microsomal desaturation and chain elongation of essential fatty acids linoleic acid (LA) and alpha-linolenic acid (LNA) to long chain polyunsaturated fatty acids (LC-PUFA). Among all trans isomers, 9t12t-18:2 isomer exerts the strongest impairment of delta6 desaturase activity [4,[12][13][14]. In adults, inhibitory effect of TFA does not cause serious undesirable health hazards when dietary LA is sufficient [15,16]. However, the fetus is very sensitive to factors that can interfere with the synthesis of LC-PUFA, because of an extensive need of arachidonic and docosahexaenoic acids for the development of neurological structures of the brain and retina [17][18][19] and for the synthesis of eicosanoids [20]. In addition, the biosynthetic activity of LC-PUFA is low in fetus [21].
As reported, the level of plasma trans octadecenoic acid was inversely related to levels of LC-PUFA in premature infants [22]. The plasma level of TFA and birth weight were also inversely related. In human fetal tissue collected after abortions at a gestational age varying from 5 to 15 weeks and umbilical venous walls from full-term neonates, there was an inverse relation between trans 18:1 isomer and the n-6 LC PUFA [23]. Weight and head circumference of full-term neonates were also inversely correlated with trans 18:1 acid in umbilical arterial vessel walls.
In this context, it has been suggested that exposure to high levels of TFA during pregnancy may impair the growth of the fetus.
The main objective of this study was to examine the incorporation of TFA into the plasma and red blood cell (RBC) lipids of the umbilical cord blood in relation with TFA composition of the maternal plasma, in order to observe a possible discriminating placental transport of the different trans isomers.
In the same way, the TFA profile of maternal plasma lipids was compared with that of the cord wall, as well as cord vein and artery phospholipids, in order to show a possible selective incorporation of the different trans isomers into lipids of cord tissues.

Subject and sample collection
The study protocol was approved by the local ethical review committee and informed consent was obtained from 97 participating pregnant women (age: 37 ± 10 years) recruited in Aquitaine between 1997 and 1999. The health state of these mothers needed a childbirth (38-42 postmenstrual weeks) by caesarean surgery.
During caesarean surgery, maternal peripheral venous blood and subcutaneous adipose tissue were removed. Cord vein blood was drawn by venipuncture from the placental portion of the umbilical cord, immediately after its clamping. Clinical data concerning the mother (age, tabbaco consumption...) and the term-born infant (height, weight, head perimeter) were registered by doctor, in a book. RBC were separated from plasma by centrifugation and umbilical cord samples were immediately rinsed with an isotonic saline solution (0.9% NaCl) and then frozen and stored at -20°C until analyzed.

Analytical procedure
Thawed samples of plasma, adipose tissue and umbilical tissues were ground in a mixture of chloroform/methanol (2:1, by vol). Total lipids were extracted from samples according to the Folch method [24]. Solvents were evaporated to dryness under a stream of nitrogen. Lipids were taken up in an appropriate volume of chloroform/methanol (2:1, by vol) and stored in glass tubes at -20°C under nitrogen. Plasma triglycerides (TG), cholesterol esters (CE) and total phospholipids (TPL) were separated by thin layer chromatography using hexane/diethyl ether/acetic acid (90:10:1, by vol).
Fatty acids methyl esters (FAME) of total lipids and fractionated lipids were prepared with 14% boron trifluoride in methanol according to Morrison and Smith [25] and stored in hexane, at -20° C under nitrogen.
Analyses of total FAME were carried out on a gas chromatograph (GC) (Carlo Erba 5160, Milano, Italy) equipped with a flame-ionization detector and a split injector. A fused-silica capillary column (BPX 70, 60 m x 0.25 mm i.d., 0.25 mum film; SGE, France) was used with H 2 as a carrier gas (inlet pressure: 90 kPa). The split ratio was 1/70. The column temperature was programmed from 150 to 200°C at 1.5°C/min, then to 230°C at 2.5° C/min and held at 230°C until completion of the analysis (30 min). The injection port and the detector were maintained at 250°C. The gas chromatographic peaks were integrated using a SP 4400 integrator (Spectra Physics, San Jose, CA).
All the results are expressed as weight percentages of fatty acids using the factor (F't) described by Wolff et al. [26]. This factor is linked to the theoretical response factor (Ft) by: F't = Ft x (fatty acid molecular weight/FAME molecular weight).

Comparison of TFA composition of mother and fetus plasma (Figure 1)
Trans fatty acid (TFA) percentage was significantly higher (p = 0.001) in maternal (0.9% of total fatty acids) than in umbilical plasma total lipids (0.6%).
This result is different from that found in a German study describing occurrence of TFA in cord plasma [22], at percentage level similar to that in the mothers. However, in that study, TFA contents of umbilical and maternal plasma total lipids were two times higher than values observed here, which was probably due to a higher intake of TFA in German women than in French, at the time of these studies.
The occurrence of TFA in cord plasma confirms the transplacental transport of the different trans isomers. Nevertheless, the prevalence of trans 18:2 isomers and more precisely of 9t12c-18:2 isomer in umbilical plasma may be explained either by a preferential transfer of this isomer across the placenta, as a consequence of sequestration process, or by an accumulation in cord plasma linked to bad metabolization of this isomer by the fetus compared with the other trans 18:2 isomers.
The placental enzyme behaviour towards trans 18:2 isomers was examined through the incorporation level of these trans diene isomers into the lipid classes of cord plasma.

TFA incorporation into different lipid classes of umbilical plasma (Figure 2)
Trans 18:2 isomers occurred in the three main umbilical plasma lipid classes, but the relative percentage values of these trans diene isomers were significantly lower (p < 0.001) in total phospholipids (TPL) (0.14% of total fatty acids) than in triglycerides (TG) (0.41%) or in cholesterol esters (CE) (0.41%).
Moreover, these three classes of plasma lipids standed out by differences in their proportions of various trans 18:2 isomers. TPL and CE were characterized by 9t12c isomer prevalence (respectively, 47 and 48% of total trans 18:2 acids), followed by the mix 9c13t+9t12t and 9c12t (respectively, 29 and 24%) in TPL and by 9c12t and the mix 9c13t+9t12t (respectively, 32 and 20%) in EC. Whereas in TG, the three trans 18:2 isomers were present in equivalent proportions (31 to 38% of total trans 18:2 acids). This finding may be explained by the metabolism of cholesterol. Fatty acid pattern of CE illustrates that of sn-2 position of phosphatidylcholine (PC), since CE are generated by activity of lecithin-cholesterol-acyl-transferase (LCAT) which transfers fatty acids from PC sn-2 position to cholesterol [27]. It is known [28] that 100% of 9t12c and 77% of 9c12t isomers are incorporated into the sn-2 position of maternal PC, whereas 75% of 9c13t+9t12t isomers mix are present in sn-1 of PC. Consequently, if it is also true for the cord plasma PC, then it could explain the preferential incorporation of 9t12c and secondly 9c12t into CE.

Incorporation of the maternal TFA isomers into phospholipids of umbilical blood and red blood cells (RBC) (Figures 3 and 4)
As shown in Figure 3, TFA percentage was significantly lower (p = 0.003) in umbilical RBC (0.25% of total fatty acids) than in umbilical plasma TPL (0.42%). Nevertheless, TFA patterns of these two compartments were similar. Trans 18:1 isomers contributed to the major portion to total TFA (55% of total TFA) in both RBC and plasma TPL. However, trans 18:2 acid level was significantly lower (p = 0.011) in RBC (0.10% of total fatty acids) than in plasma TPL (0.14%). Moreover, trans 18:2 acid profiles ( Figure 4) were different in these two compartments. Although the three trans 18:2 isomer proportions were equivalent (37 to 40% of total trans 18:2 acids) in TPL of RBC, 9t12c isomer predominated (47%) over 9c13t + 9t12t isomer mix (29%) and 9c12t isomer (24%) in plasma TPL.

Maternal TFA incorporation into phospholipids of cord and vessel walls (Figure 5)
After having investigated TFA incorporation into both plasma and RBC lipids of cord vein blood, our attention was then paid to the TFA incorporation into TPL of umbilical cord parietal tissue and umbilical vessels.
Except trans 16:1 acids, all trans isomers already seen in umbilical plasma were detected in both parietal and vessel TPL of umbilical cord. TFA content, all isomers confounded, of cord wall TPL (0.31% of total fatty acids) didn't differ significantly from that found in cord vein (0.28%) or artery TPL (0.35%). Nevertheless, proportions of trans 18:1 and 18:2 isomers were similar (respectively, 48 and 52% of total fatty acids) in cord wall and vein TPL, whereas in artery TPL, trans 18:2 isomers predominated (68% of total TFA) significantly (p = 0.001) over trans 18:1 isomers (32% of total TFA). Furthermore, these three compartments of umbilical cord (parietal tissue, vein and artery) were characterized by prevalence of 9c13t + 9t12t isomers compared with the other trans 18:2 isomers. Nevertheless, the isomer mix level was significantly higher (p < 0.001) in artery TPL (0.17% of total fatty acids) than in vein TPL (0.09%), itself significantly higher (p = 0.005) than in cord wall TPL (0.06%).
Moreover, in artery TPL, percentage value of 9c13t + 9t12t isomer mix was inversely correlated (r = -0.703, p = 0.003) with arachidonic acid (AA) content, but not with that of linoleic acid (LA), precursor of AA. This result is in agreement with the inverse relation observed between TFA and LC-PUFA in umbilical plasma lipids of both healthy children [29] and preterm infant [22]. This inverse relation was also verified between trans 18:1 isomer and the n-6 LC PUFA in human fetal tissue collected after abortions (gestational age 5-15 weeks) [23].
If we admit that fatty acid pattern of umbilical artery TPL reflects that of fetal tissue TPL, then these results could demonstrate a potential impairment by 9c13t + 9t12t isomer mix of LA metabolism, or preferential uptake of maternal plasma AA by the placenta. Major part of LC-PUFA found in fetal tissues comes from maternal blood, because placental delta6 desaturase activity is too poor to provide sufficiently LC-PUFA for fetus development [30]. Nevertheless, concerning AA, selective transplacental transport is supposed to be less important than fetal synthesis [31], thus interference of 9c13t + 9t12t isomer mix with delta6 desaturase activity could be the best explanation to our results. Insofar as AA is of great physiological importance during fetal development as essential membrane component of neural tissues and as precursor for eicosanoid synthesis, it could be assumed that accumulation of these trans dienes might have side effects on fetus development. However, these results must be considered with a lot of caution, since it concerns the mix of different trans 18:2 isomers.
Some studies [22,23] have suggested that exposure to high levels of TFA during pregnancy may affect intrauterine human growth, in particular birth weight and head circumference.
So, after examination of TFA incorporation into umbilical plasma lipids and TPL of cord and vessel walls, we have devoted us to assess, for these French pregnant women, potential impact of their TFA intake during pregnancy on either weight or head circumference of newborn. (Figures 6 to 9) Because TFA content of adipose tissue is a reliable biochemical indicator of the usual TFA intake level [32,33], we studied newborn anthropometric parameters in relation to TFA level of maternal adipose tissue lipids (Figures 6 and 7). Based on dietary questionnaire, we found that average TFA consumption of this pregnant women population was 3.1 ± 1.2g/d/pers., ranging from 0.9 to 6.5g/d/pers. (or 0.5 to 2.5% of total energy intake). No correlation was observed between TFA level of maternal adipose tissue and either newborn weight (Figure 6) or head circumference (Figure 7). Correlation absence persisted even after having applied corrections which took account of confounding factors such as mother's age, BMI and possible tabacco consumption.

Relation between newborn anthropometric parameters and TFA levels in maternal and umbilical plasma
This first result shows that TFA intake level of these French pregnant women seems to be sufficiently low to induce no growth failure.
However, we have verified in the second time if TFA amount, transferred across the placenta from maternal plasma lipids, could be put in relation to previous antropometric parameters (Figures 8 and  9). As for the adipose tissue, we noticed that TFA level of maternal plasma lipids was related to neither newborn weight (Figure 8) nor head circumference (Figure 9), even after multifactorial corrections.
Our results do not agree with those reported by Koletzko et al. [22] who demonstrated negative correlations between birth weight of premature infant and percentage of either total TFA in umbilical TPL or trans 18:1 acid percentage in CE. Similarly, Van Houvelingen et al. [23] have also observed inverse relation between trans 18:1 acid in umbilical arterial vessel walls and weight and head circumference of full-term neonates.
This discrepancy might be explained by sensitivity to TFA in full-term lower than in preterm infant whose LC-PUFA stores are extremely more limited compared to term infant [29], and by TFA intake of mothers which was higher in Germany and Netherlands than in Aquitaine [34].

CONCLUSION
In summary, we observed differences in TFA proportions between maternal and cord plasma total lipids. These data could demonstrate either potential discriminating placental transport of different TFA isomers or metabolism selectivity of different TFA isomers by the fetus. Trans isomers were incorporated not only into lipids of cord blood (plasma and RBC) afferent to fetus but also into TPL of parietal tissue and vessels of umbilical cord. Moreover, inverse correlation found between 9c13t + 9t12t isomer mix and AA levels in artery TPL suggested that these trans 18:2 isomers might enhance LC-PUFA deficiency of fetus. However, this hypothese must be supported by further investigations since it concerns mix of different trans 18:2 isomers.
Finally, TFA intake level of these French pregnant women, assessed to 3.1 gram per day, seems to be sufficiently low not to affect negatively weight and head circumference of newborn.