Fatty acid composition of the milk lipids of Nepalese women: correlation between fatty acid composition of serum phospholipids and melting point

Fatty acid composition of the milk lipids of Nepalese women: correlation between fatty acid composition of serum phospholipids and melting point

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3),147^156 & 2001 Harcourt Publishers Ltd doi:10.1054/plef.2001.0303, available online...

379KB Sizes 0 Downloads 99 Views

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3),147^156 & 2001 Harcourt Publishers Ltd doi:10.1054/plef.2001.0303, available online at http://www.idealibrary.com on

Fatty acid composition of the milk lipids of Nepalese women: correlation between fatty acid composition of serum phospholipids and melting point R. H. Glew,1 Y-S. Huang,2 T. A.Vander Jagt,1 L-T. Chuang,2 S. K. Bhatt,3 M. A. Magnussen,3 D. J.VanderJagt1 1 Department of Biochemistry and Molecular Biology, University of New Mexico School of Medicine, Albuquerque, NM, USA; 2Ross Products Division, Abbott Laboratories, Columbus, OH, USA; 3Public Health Clinic, Kailali District, Dhangadhi, Nepal

Summary Milk was collected from 36 Nepalese women,15 to 32 years of age, in order to investigate relationships between the proportions of intermediate chain-length (C10^C14) fatty acids and critical n-3 and n-6 polyunsaturated fatty acids in the milk lipids they were producing. Serum was also obtained from these lactating women and the fatty acid composition of their serum phospholipid fraction was determined and compared with that of the corresponding milk lipid fraction. Compared to women in technologically advanced parts of the world, the serum phospholipids of the Nepalese women contained nutritionally adequate proportionsoflinoleic acid (LA) (16.8%), a-linolenic acid (ALA) (0.53%), arachidonic acid (AA) (5.69%), and docosahexaenoic acid (DHA) (1.42%). However, although the milk lipids contained adequate proportions of ALA (1.81%), AA (0.43%), and DHA (0.23%), the lipids contained low to moderate percentages of LA (mean, 9.05%). Positive correlations were observed between the proportions of AA (P = 0.001, r = 0.50) and ALA (P = 0.03, r = 0.36) in the serum phospholipids and milk lipids of the women. As the proportion of C10^Cl4 fatty acids in the milk lipids increased from10% to 40%, there was preferential retention of three critical n-3 and n-6 fattyacids (ALA, AA, and DHA) at the expense of two relativelyabundant nonessential fattyacids, namely stearic acid and oleic acid. In addition, using fatty acid melting point data and the mol fraction of the 9 most abundant fatty acids in the milk, we estimated the mean melting point (MMP) of the milk lipids of the Nepalese women.The MMPs ranged from 29.3 to 40.58C (median, 35.58C). These resultsindicate that:1) the levels of AA and ALA in the blood of lactating mothersinfluence the levels of these fattyacidsin the milk they produce; 2) when the mammary gland produces a milk that is rich in C10^Cl4 fatty acids, it somehow regulates triglyceride synthesis in such a way as to ensure that the milk will provide the exclusively breast-fed infant with the amounts of the critical n-3 and n-6 fatty acids it requires for normal growth and development; and 3) the melting point of the milk lipid fraction is determined mainly by the mol % of the intermediate chain-length (C10^C14) fatty acids, oleic acid, linoleic acid, and a-linolenic acid. & 2001Harcourt Publishers Ltd

INTRODUCTION

Received 5 June 2001 Accepted 19 July 2001 Correspondence to: Dorothy J.VanderJagt, PhD, Department of Biochemistry and Molecular Biology,Room 249 BMSB, School of Medicine,University of New Mexico, Albuquerque, NM 87131, USA.Tel.: (505) 272-5799; Fax: (505) 272-6587; E-mail: [email protected] Source of support : This study was supported by a Minority International ResearchTraining (MIRT) grant from the Fogarty International Center of the National Institutes of Health.

& 2001Harcourt Publishers Ltd

Human milk, unlike cow milk, contains relatively large amounts of intermediate chain-length fatty acids. These 10 to 14-carbon, saturated fatty acids are sometimes referred to as de novo fatty acids because they are synthesized from acetyl-CoA in mammary tissue, as opposed to the long-chain fatty acids in milk which are comprised of 16–22-carbon atoms and which are

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

148

Glew et al.

acquired from the diet, plasma lipoproteins or adipose stores.1 The proportion of de novo fatty acids in the triglyceride fraction of human milk varies greatly within populations and across different geographical regions of the world. Lactating women whose diets are high in carbohydrates tend to produce a milk fat that is rich in C10–C14 fatty acids.2,3 Until recently it was widely accepted that the main function of intermediate chain-length fatty acids was to provide the breast-fed infant with a fuel that may be readily digested, absorbed, and rapidly oxidized, especially during the critical first few months of life when the infant’s ability to produce and secrete digestive tract lipases, synthesize chylomicrons, and transport longchain fatty acids into mitochondria in a carnitinedependent manner may not be sufficiently developed. In a recent study of the fatty acid composition of the milk of Fulani women conducted in northern Nigeria,4 we discovered a novel phenomenon that revealed a possible, but heretofore unrecognized, role for C10–C14 fatty acids in human milk. We found that among the milk specimens of the 34 Fulani women we studied, as the percentage of C10–Cl4 fatty acids in the milk lipid fraction of that population increased from 15 to 65%, there was not the same proportional decrease in the percentages of most of the nonessential fatty acids such as oleic acid and stearic acid or in the proportions of three critical n-3 and n-6 fatty acids which included one of the essential fatty acids (a-linolenic acid), and arachidonic acid and docosahexaenoic acid. For example, we observed that as the de novo fatty acid content of the milk fat of the Fulani women increased 4.3-fold, the proportions of oleic acid and stearic acid decreased by 86% and 69%, respectively; in contrast, over the same-fold increase in C10–Cl4 fatty acids, the proportions of a-linolenic acid and arachidonic acid decreased by only 2.3% and 28%, respectively. These observations led us to speculate that de novo fatty acids may regulate the synthesis and composition of triglycerides in human milk; more specifically, we hypothesized that in humans C10–-Cl4 fatty acids cause mammary tissue to synthesize triglyceride molecules that are enriched in particular critical n-3 and n-6 fatty acids at the expense of the nonessential fatty acids. The mechanism that accounts for this phenomenon is obscure. The results of a subsequent study of the fatty acid composition of the milk of Kanuri women in northern Nigeria5 were in agreement with our observations regarding the relationships between C10–Cl4 fatty acids and n-3 and n-6 fatty acids in the milk of Fulani women.4 Out of concern about the generality and universality of our two studies which pointed to an apparent close relationship between the proportions of intermediate chain-length fatty acids and essential fatty acids in human milk, we elected to investigate this question in a

population of lactating women outside of sub-Saharan Africa. We therefore collected milk from native Nepalese women and analyzed it for its fatty acid composition using the very same chemical and statistical methods of analysis that we employed in our two previous studies that were conducted in Nigeria. Our aims in this regard were two-fold: firstly, to test the de novo fatty acid/ preferential retention of essential fatty acids’ hypothesis; and secondly, to extend our previous study of lactating Nepalese women,6 which involved only their milk, in such a way that we might investigate the relationships between the fatty acid composition of the serum phospholipids of the women and that of the milk fat they were producing. A third aim of the present study was to estimate the mean melting points of the lipid fractions of the Nepalese women. Jensen and Patton7 demonstrated recently that one can accurately calculate the mean melting point (MMP) of human milk triglycerides from the mol fractions of the most abundant fatty acids and their respective melting points. They showed that for a number of populations around the world, the MMP of the lipid fraction of human milk was lower than normal body temperature (378C), and hypothesized that the mammary gland regulates its acyltransferase system in such a way as to ensure that the triglycerides that are produced remain liquid at body temperature. Specifically, we wanted to test Jensen and Patton’s hypothesis in Nepalese women and, more importantly, to determine which fatty acids were the most important in determining the MMP of the lipid fraction of human milk.

SUBJECTS AND METHODS

Subjects The study population consisted of 36 Nepalese women who were recruited from the village of Dhangadhi, in rural Nepal. Milk was obtained during the months of January through April 1999, and the lactation stage of the women ranged from 2 weeks to 6 months (mean, 2.9 months) postpartum. The milk was aliquoted into 1.5 ml cryovials and stored at 7208C for 3–6 weeks until the samples were transported on dry ice to the United States for analysis. The Ethics Review Board of the Ministry of Health in Nepal approved this study.

Fatty Acid Analysis Milk samples were thawed, warmed to 378C and vortexed vigorously before analysis. Lipids were extracted using the method described by Folch et al.8 Boron trifluoridemethanol (14%, w/v) was used for transesterfication of

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

& 2001Harcourt Publishers Ltd

Fatty acid composition of milk lipids

the total lipid fractions and the fatty acid methyl esters were extracted using hexane.9 Fatty acids were quantified using a Hewlett Packard Gas Chromatograph (5890 Series II; Mississauga, Ontario) equipped with a flame-ionization detector and a 50 m fused silica capillary column coated with 0.25 m Supelcowax 10 (Supelco Inc., Bellefonte, PA).10 The injector and detector temperatures were 2308C. Commercial standards (Nu-Chek-Prep Inc., Elysian, MN, USA and Supelco Inc., Bellefonte, PA) were used for identification and quantification of fatty acid methyl esters. Results are expressed as mg/100 g total fatty acid.

Table 1 Anthropometric characteristics of the study population Mean (+SD) Number of subjects Age (yr) Height (cm) Weight (kg) BMI (kg/m2) Parity

Statistical Analysis Two-sample t-tests and correlation analysis were performed using NCSS 6.0 (Kaysville, Utah). A P-value of 0.05 was considered significant. RESULTS

Comments on the study population The average age of the 36 women who participated in this study was 22.2 years and the mean age of the infants they were breastfeeding was 2.9 months (Table 1). With regard to the period of lactation, the women from whom milk was collected had been lactating for 2 to 26 weeks. The mean body mass index of the Nepalese women was 20.9 kg/m2.

Fatty acid composition of serum phospholipids Table 2 summarizes the fatty acid composition of the serum phospholipids of the Nepalese subjects. It has been established that the fatty acid profile of serum phospholipids provides a reliable means of assessing the nutrient status of an individual with respect to essential fatty acids and critical n-3 and n-6 polyunsaturated fatty acids.10–12 & 2001Harcourt Publishers Ltd

36 22.2 (+3.2) 152.6 (+7.42) 48.8 (+5.9) 20.9 (+2.6) 2.5 (+1.4)

Table 2 The fatty acid composition (wt %) of total lipids and serum phospholipids of Nepalese women

Estimation of the melting points of human milk lipids We used the method described by Jensen and Patton7 to calculate the melting point (MP) of the lipid fraction of the human milk specimens from fatty acid composition data. First, we determined the mol % of each of the 9 most abundant fatty acids. This was accomplished by dividing the mass % of each fatty acid obtained by gas-liquid chromatography of fatty acid methyl esters by the respective molecular weights of the fatty acids and then calculating the mol % of each. Next, we multiplied the mol % by the [MPþ100] of each fatty acid to obtain the MP fractions. Finally, the MP fractions were summed and 100 subtracted from the sum to provide the estimate of the lipid’s mean melting point (MMP).

149

Saturated C10:0 C12:0 C14:0 Total de novo C15:0 C16:0 C18:0 C20:0 C22:0 Monounsaturated C16:1n-7 C1 8:1n-9 C18:1n-7 C18:1n-5 C20:1n-9 C20:1n-7 C24:1 Polyunsaturated (n-6) C:18:2 C18:3 C20:2 C20:3 C20:4 C22:4 C22:5 (n-3) C18:3 C20:5 C22:5 C22:6

Blood serum phospholipids Mean (SD) (n=36)

Milk triglycerides Mean (SD) (n=36)

N/D N/D 0.58 (0.15) 0.58 (0.15) 0.50 (0.12) 30.8 (1.76) 14.4 (1.48) 0.52 (0.15) 0.59 (0.21)

2.48 (0.76) 10.4 (4.09) 10.9 (3.83) 23.7 (8.39) 0.50 (0.22) 22.1 (3.42) 5.01 (1.68) 0.21 (0.07) 0.09 (0.03)

1.31 (0.47) 12.7 (2.37) 1.99 (0.43) 0.20 (0.05) 0.74 (0.28) 0.22 (0.11) 3.26 (1.26)

3.34 (1.00) 24.3 (4.45) 2.40 (0.61) 0.16 (0.09) 1.61 (0.84) 0.23 (0.11) 0.61 (0.27)

16.8 (2.70) 0.24 (0.08) 0.62 (0.15) 2.65 (0.44) 5.69 (0.92) 0.37 (0.14) 0.31 (0.09)

9.05 (2.49) 0.13 (0.04) 0.32 (0.09) 0.39 (0.12) 0.43 (0.09) 0.10 (0.02) 0.05 (0.02)

0.53 (0.17) 1.11 (0.35) 0.93 (0.25) 1.42 (0.44)

1.81 (0.77) 0.15 (0.04) 0.24 (0.09) 0.23 (0.06)

N/D, not detected.

The proportions of the two essential fatty acids, linoleic acid (16.8%) and a-linolenic acid (0.53%), in the serum phospholipid fraction of the Nepalese women, either exceeded or were within the range of values reported for other populations,13–16 thereby indicating that the nutrition of the subjects in this study was adequate with respect to essential fatty acids. Similarly, the proportions of arachidonic acid (5.69%) and docosahexaenoic acid (1.42%) were within the range of values reported by other investigators for adults.13–17

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

150

Glew et al.

Fatty acid composition of milk lipids Since triglycerides account for 98% or more of the lipids in human milk,18 the fatty acid composition of the total lipid fraction of milk reflects that of the triglycerides in the milk. As shown in Table 2, the mean fatty acid composition of the milk lipids of the 36 Nepalese women who participated in this study was nearly identical to that of the 48 women in Nepal that we studied one year earlier.6 In the present study, intermediate fatty acids accounted for 23.7% of the fatty acid total. The proportions of the critical n-3 and n-6 fatty acids were as follows: linoleic acid, 9.05%; a-linolenic acid, 1.81%; arachidonic acid, 0.43%; and docosahexaenoic acid, 0.23%. Except for the proportion of linoleic acid which was relatively low compared to what has been observed in other studies of human milk elsewhere in the world,3,4–6,19–26 the proportions of the other three polyunsaturated fatty acids were as high, or higher, than those reported by other investigators. When we conducted a statistical analysis to inquire about possible relationships between the proportions of individual fatty acids in the milk specimens and the mother’s age, parity, BMI, or time post-gestation, no statistically significant correlations were observed.

Effect of nutritional status Does the nutritional status of the mother with respect to critical n-3 and n-6 fatty acids influence the content of these same fatty acids in the milk she produces? To address this question, we analyzed the extent of the correlation between the proportions of linoleic acid, alinolenic acid, arachidonic acid and docosahexanoic acid in the serum phospholipid fraction of the lactating women with those of their total milk lipids: statistically significant correlations were observed, but only for the n-3 fatty acid a-linolenic acid (Fig. 1, panel A) and the n-6 fatty acid arachidonic acid (Fig. 1, panel B).

Proportions of C10^C14, essential, and nonessential fatty acids in the milk of Nepalese women We have previously observed what we have referred to as a conservational effect of de novo fatty acids on the content of essential n-3 and n-6 fatty acids in human milk lipids.4, 5 To determine if this intriguing phenomenon were manifest in the milk of Nepalese women, we plotted the proportions of various nonessential fatty acids (palmitic acid, stearic acid, oleic acid) and four critical n-3 fatty acids (linoleic acid, a-linolenic acid, arachidonic acid, docosahexanoic acid) versus the proportions of C10–C14 fatty acids in the milk lipid fraction. The results of this analysis are shown in Figures 2 and 3 and

Fig. 1 (A) The relationship between the percentage of a-linolenic acid (C18: 3n-3) in the maternalserum phospholipids and the percentage of Cl8: 3n-3 in the milk lipids: Y=70.029X þ 2.51, P = 0.06, r = 0.32. (B) The relationship between the percentage of arachidonic acid (C20: 4n-6) in the maternal serum phospholipids and the percentage of C20: 4n-6 in the milk lipids: Y=0.004Xþ0.339, P = 0.01, r = 0.39.

summarized in Table 3. As shown in Figure 2, as the percentage of C10–Cl4 fatty acids in the milk lipid fraction increased four-fold from 10% to 40%, there was a decline of 18% and 46%, respectively, in the proportions of the nonessential fatty acids palmitic acid (panel A) and oleic acid (panel B). Over the same-fold increase in the percentage of C10–C14 fatty acids, the proportion of another nonessential fatty acid, namely stearic acid, decreased by 60% (Table 3). As shown in Figure 3, with regard to the nutritionally important n-3 and n-6 fatty acids, three of them either decreased only slightly (panel A, linoleic acid, 12%) or actually exhibited enrichment (panel C, arachidonic acid, 30%; panel D, docosahexaenoic acid, 71%) over the 4-fold increase in the proportion of de novo fatty acids. Noteworthy is the fact that a-linolenic acid (Fig. 3, panel B) appeared to behave more like the nonessential fatty acids in that as the contribution

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

& 2001Harcourt Publishers Ltd

Fatty acid composition of milk lipids

Fig. 2 (A) The relationship between the percentage of palmitic acid (C16: 0) in milk triglycerides and the percentage of total C10^C14 fatty acids: Y =70.149X þ 25.6, P = 0.03, r = 0.36. (B) The relationship between the percentage of oleic acid (C18:0) in milk triglycerides and the percentage of total C10^C14 fatty acids: Y=70.137X þ8.26, P50.001, r = 0.68.

C10–Cl4 acids made to the fatty acid composition of the milk lipid fraction increased from 10% to 40%, the percentage of this essential n-3 fatty acid declined by 40%. When we averaged the changes that occurred in the three nonessential fatty acids and the four essential fatty acids as the percentage of C10–C14 fatty acids increased from 10% to 40% (Table 3), we found that for the four n-3 and n-6 fatty acids, there was an overall increase of 12% in their representation in the lipid fraction of the milk and a concomitant 43% decrease in the proportion of the three nonessential fatty acids.

determinants of MMP. To answer this question, we compared the strength of the correlation obtained when the MMP of each of the 36 milk lipid fractions from different women was plotted versus the mol % of the individual fatty acids used to calculate the MMP. Qualitatively, the various plots fell into one of three different patterns; ones where: 1) there was no correlation between MMP and mol %; 2) the plot was linear with a positive slope; or 3) the plot was linear with a negative slope. The strongest positive MMP versus mol % correlation and the one that had the steepest slope was seen with the saturated C-10 intermediate chain-length fatty acid, decanoic acid: the slope was þ8.58C/0.1 mol % and the r value was 0.34 (Table 4) (Fig. 4). To our surprise, the mol % of palmitic acid, the most abundant long-chain saturated fatty acid in all of the milk lipid fractions (Table 4), did not correlate significantly with MMP (Fig. 5); that is, the slope of the plot of MMP versus mol % for palmitic acid was not significantly different from zero. The same was true for stearic acid (Fig. 6). Thus, the saturated fatty acid that seemed to have the greatest influence on the MMP of the milk lipids was decanoic acid (Fig. 4). This conclusion is supported by the data in Table 4 where we have summarized the values for the MMP versus mol % plots of the nine most abundant fatty acids upon which we based our MMP estimates, and the melting point coefficients, where the melting point coefficient represents the slope of the MMP versus mol % plots like those shown in Figures 4–7. The expectation that the MMP would correlate negatively with the mol % of one or more of the unsaturated fatty acids was borne out by the data shown in Figure 7. The absolute magnitude of the slopes of the MMP versus mol % plots for oleic acid, linoleic acid, and a-linolenic acid were 3.3, 7.9, and 24.98C/0.1 mol %, respectively. Furthermore, the r values of these plots increased progressively from oleic acid (0.65), linoleic acid (0.73), to a-linolenic acid (0.79), Interestingly, arachidonic acid, with its four carbon-carbon double bonds and very low MP (7508C), showed no MMP versus mol % correlation (data not shown). Thus, on a mol % basis, the most potent unsaturated fatty acid in terms of reducing the MMP of the lipid fraction of the milk specimens was a-linolenic acid. DISCUSSION

Estimation of the mean melting points (MMPs) of the lipid fraction of the milk The MMPs of the lipid fraction of the 36 human milk specimens ranged from 29.3 to 40.58C and the median MMP was 35.58C. The main reason for performing this kind of analysis was to determine which fatty acids were the major & 2001Harcourt Publishers Ltd

151

In our previous study of the fatty acid composition of the milk lipids of women in Nepal,6 we found that, relative to the milk of women in most other parts of the world, the Nepalese women were producing a milk fat that contained a low proportion of linoleic acid (7.91%), the essential n-6 fatty acid. However, in that initial study of

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

152

Glew et al.

Fig. 3 (A) The relationship between the percentage of linoleic acid (C18: 2n=6) in milk triglycerides and the percentage of total C10^C14 fatty acids: Y=70.037Xþ9.94, P = 0.46, r = 0.10. (B) The relationship between the percentage of a-linolenic acid in milk triglycerides and the percentage of total C10^C14 fatty acids: Y=70.029Xþ2.51, P = 0.06, r = 0.32. (C) The relationship between the percentage of arachidonic acid (C20:4n-6) in milk triglycerides and the percentage oftotal C10^Cl4 fattyacids:Y=0.0043Xþ0.339, P = 0.01, r = 0.39. (D) The relationship between the percentage of docosohexaenoic acid (C22: 6n-3) in milk triglycerides and the percentage of total C10^Cl4 fatty acids: Y = 0.004Xþ0.129, P50.001, r=0.57.

Table 3 Analysis of the correlation between the proportions of various essential and nonessential fatty acids and the proportions of C10^C14 (de novo) fatty acids in the milk triglycerides of the Nepalese women Fatty acid Nonesseptial C16:0 C18:0 C18:1n-9 Essential C18:2n-6 C18:3n-3 C20:4n-6 C22:6n-3

P value

r

0.02 50.001 50.001

0.36 0.67 0.88

0.82 0.40 0.54

NS 0.05 0.01 50.001

0.10 0.32 0.39 0.57

0.88 0.60 1.30 1.71

% at 40% de novo %at 10% de novo

NS, not significant

lactating Nepalese women, we remarked that their milk fat contained a relatively high proportion of a-linolenic acid (1.93%) and adequate percentages of arachidonic acid (0.35%) and docosahexaenoic acid (0.21%). The results of the present study agree well with those of our previous study; in the present study, we found that the

Table 4 Summaryof melting point coefficients for various fattyacids in milk lipids of Nepalese, expressed as 8C per 0.1 mol % of fattyacids* Fatty acid

Melting point coefficient* (8C per 0.1 mol %)

r value

P value

C10:0 C12:0 C14:0 C16:0 C18:0 C18:1 C18:2 C18:3 C20:4n-6

þ8.5 þ2.5 þ4.4 NS NS 73.3 77.9 724.9 NS

0.34 0.46 0.67 0.20 0.01 0.65 0.73 0.79 0.14

0.043 0.004 50.001 ^ ^ 50.001 50.001 50.001 ^

NS,The slope of the plot of mean melting point versus mol % for16:0, 18:0 and 20:4n-6 wasnot significantlydifferent from +1.08C per 0.1 mol %.*The melting point coefficient was determined from the slope ofthe plot of mean melting point (8C) versus mol % of the particular fatty acid.

milk of the Nepalese women contained 9.05% linoleic acid, 1.81% a-linolenic acid, 0.43% arachidonic acid, and 0.23% docosahexaenoic acid. With regard to the MMP analysis component of the present study, two main findings emerged. First, the

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

& 2001Harcourt Publishers Ltd

Fatty acid composition of milk lipids

153

Fig. 4 The relationship between the mean melting point (MMP) of milk lipidsand the mol % of decanoic acid (C10:0):Y = 85.1Xþ32.8, P = 0.04, r = 0.33.

Fig. 6 The relationship between the mean melting point (MMP) of milk lipids and the mol % of stearic acid (C18:0): Y = 733.3Xþ44.3, P =N.S., p=0.001

Fig. 5 The relationship between the mean melting point (MMP) of milk lipids and the mol % of palmitic acid (C16:0): Y = 15.1Xþ32.5, P = 0.23, r = 0.20.

Fig. 7 The relationship between the mean melting point and the mol % of oleic acid (C18:1n-9) P =50.001 (*): Yþ^33.3Xþ 44.3, r = 0.65; linoleic acid (C18:2 n-6) (*): Y=7249X þ 40.8, P50.001, r = 0.73; and a-linolenic (Cl8:3 n-3) (!): Y=779.15Xþ43.4, P50.001, r = 0.79.

median MMP of the lipid fractions of the milk of the Nepalese women was 35.58C, which is several degrees below normal body temperature (378C). Thus, in general, the melting points of the milk fat of these women support Jensen and Patton’s contention7 that the liquidity of the triglycerides in human milk is maintained by the mammary gland’s ability to synthesize triglycerides with MPs that are lower than a woman’s normal body temperature. In the case of the milk lipid specimens in the present study, we found that 4 out of 36 samples had MMPs greater than 378C. Jensen and Patton7 would argue that small quantities of high-melting-point triglycerides would be dissolved in those triglycerides whose MPs were below 378C. The second major finding of the MMP study was the identification of those fatty acids that appear to be the

main determinants of the MMP of a particular milk lipid fraction. The intermediate chain-length fatty acids, decanoic acid, lauric acid and myristic acid seemed to be much more important than the long-chain fatty acids palmitic acid and stearic acid in elevating the MP of the lipid fraction of the milk of the Nepalese women (Table 4). Conversely, linoleic acid and a-linolenic acid appear to serve as the fatty acids which ensure that the milk triglycerides will have a fluid character in lactating women. It remains to be seen how the mammary gland regulates the fatty acid composition of the triglycerides it synthesizes so as to ensure that these lipids, which account for 496% of the total lipids in human milk,3 remain fluid and transportable across cellular membranes into the lumen of the alveoli of the breast. As suggested

& 2001Harcourt Publishers Ltd

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

154

Glew et al.

by Jensen and Patton,7 it is likely that the fatty acyltransferases in the mammary gland play an important role in this regulatory process. It will be interesting to see if the generalizations reached in the present study about which fatty acids seem to be most important in regulating the MMP of human milk lipids are supported by similar studies in other ethnic groups where the lactating women consume very different diets and where the fatty acid compositions of the milk triglyceride fractions vary greatly. We speculated in an earlier report6 that the relatively low amount of linoleic acid in the milk of Nepalese women might have been reflective of the suboptimum linoleic acid nutrition of these women. This explanation seems unlikely, however, in light of the results of our analysis of the fatty acid composition of the serum phospholipids of lactating Nepalese women (Table 2). The contribution a particular essential fatty acid makes to the fatty acid composition of an individual’s serum phospholipids is widely regarded as a measure of the nutritional status of that individual with respect to the essential fatty acid in question.10–12 In healthy adults and children, linoleic acid constitutes 12.8–24.1% 13–15 of the total fatty acids in the phospholipid fraction of serum. Thus, the fact that the mean proportion of linoleic acid in the serum phospholipid fraction of the Nepalese women was 16.8% (Table 2) suggests that their diets were probably providing them with adequate amounts of this essential n-6 fatty acid. There must, therefore, be some other reason(s) why the milk the Nepalese women were producing in the previous study and the present one contained relatively small amounts of linoleic acid. Perhaps the fatty acid transport system in the mammary gland of the Nepalese women and which functions to extract fatty acids from circulating lipoproteins was somehow discriminating against linoleic acid. There is a precedent for such selective uptake of fatty acids by tissues. For example, the placenta selectively extracts arachidonic acid and DHA from the maternal circulation and transports it into the fetal compartment, thereby enriching these long-chain polyunsaturated fatty acids (LCPUFA) in the fetus during the third trimester when fetal demands for vascular and neural growth are greatest.14 The preferential uptake of LCPUFA by a plasma membrane fatty acid-binding protein located in the microsomal villous membrane ensures that LCPUFA are more concentrated in the fetus than in the mother.27 Alternatively, it might be that the acyltransferases in the mammary cells of the Nepalese women which are responsible for synthesizing the triglycerides of milk were utilizing other fatty acids in preference to linoleic acid. Another result of the present study was the finding of a statistically significant correlation between the proportion of certain critical fatty acids in the serum phospho-

lipids of the lactating Nepalese women and the percentage of the same fatty acid in their milk lipids. Specifically, when we plotted the percentage of a-linolenic acid in the milk lipids of Nepalese women versus the percentage of a-linolenic acid in their serum phospholipids (Fig. 1A), a statistically significant (P = 0.03) positive correlation (r = 0.36) was obtained. Likewise, when the same type of plot was constructed for arachidonic acid (Fig. lB), again, a statistically significant (P50.001) positive correlation (r = 0.50) was found. These results indicate that the levels of a-linolenic acid and arachidonic acid in the circulation of these women are related to the levels of these fatty acids in the milk they produce. More generally, these data support the contention articulated elsewhere28 that the nutritional status of a lactating woman with respect to these two n-3 and n-6 fatty acids has a direct bearing on the amounts of these critical fatty acids in the milk she produces. The literature contains few studies that have examined the issue of the influence of essential fatty acid status of lactating women on the fatty acid composition of the triglyceride fraction of the milk they produce. However, we recently completed such a study among nursing Fulani women in northern Nigeria.28 The Fulani are seminomadic pastoralists whose livelihood is centered around cattle and dairy products. In that study, which involved approximately the same number of women as were enrolled in the present study of Nepalese women, we found positive and highly significant correlations between the percentages of linoleic acid, a-linoleic acid, arachidonic acid and docosahexaenoic acid in the serum phospholipids of Fulani women and the percentages of these same four n-3 and n-6 fatty acids in their milk lipids. It is puzzling why in a study conducted with Fulani women by VanderJagt and coworkers28 statistically significant correlations were observed between the serum phospholipid and milk lipid fractions of the lactating women for linoleic acid and docosahexaenoic acid but not in the present study of Nepalese women. VanderJagt and colleagues’ finding28 of a remarkably strong correlation between the proportion of docosahexaenoic acid in the sera of Fulani women and that of their milk (P50.001, r = 0.93) could be explained by the fact that the Fulani women they studied had a higher proportion of DHA in their serum phospholipids than did the Nepalese women who participated in the present study (means, 1.97% versus 1.42%). However, the same explanation cannot account for our failure to observe a statistically significant relationship between the a-linolenic acid percentages in the serum phospholipids and milk lipids of the Nepalese women, since the serum phospholipids of the Fulani women in our previous study contained much less alinolenic acid than those of the Nepalese women (0.20% versus 0.53%) in the present study.

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

& 2001Harcourt Publishers Ltd

Fatty acid composition of milk lipids

The data from this study also lend further support to the hypothesis that intermediate chain-length fatty acids, by some as-yet unknown mechanism, cause the mammary gland to synthesize triglycerides that are enriched for three critical n-3 and n-6 fatty acids, namely alinolenic acid, arachidonic acid, and DHA.4,5,28 For example, we found in our previous study of the fatty acid composition of the milk of various ethnic groups in Nigeria4 that as the proportion of C10–C14 fatty acids in the milk fat increased from 10–65%, the percentage of alinolenic in the milk fat decreased by only 2.3%, whereas the proportion of stearic acid in the milk fat decreased by 87%. In the present report, we found that as the contribution of de novo fatty acids in the milk fat of the Nepalese women increased from 10% to 40%, the proportions of the nonessential fatty acids palmitate, stearate, and oleate all decreased, whereas the percentage of linoleic acid remained constant and that of arachidonic acid and DHA actually increased (Fig. 3, A–D). Several biochemical mechanisms that could account for this differential effect of C10–Cl4 fatty acids on nonessential versus critical n-3 and n-6 fatty acids in human milk fat are discussed elsewhere.4 In conclusion, this study has confirmed that, with the exception of linoleic acid, the milk fat produced by Nepalese women appears to contain adequate amounts of DHA and arachidonic acid, and supra-normal levels of alinolenic acid. The relatively low proportions of linoleic acid in the milk fat of the Nepalese subjects in this study cannot be explained by inadequate linoleic acid nutrition in these lactating women since the mean percentage of linoleic acid in their serum phospholipids was well within the range of values reported for women and children elsewhere in the world. It will be interesting to see if the relatively low amounts of linoleic acid we found in the milk of the Nepalese women who participated in this study and our previous one6 result in correspondingly low levels of this essential n-6 fatty acids in the serum phospholipids of the infants who are exclusively breastfed by these women. Such a study is in progress.

REFERENCES 1. Odle J. New insights into the utilization of medium-chain triglycerides by the neonate: observations from a piglet model. J Nutr 1997; 127: 1061–1067. 2. Read W. W., Lutz P. G., Tashjian A. Human milk lipids: the influence of dietary carbohydrates and fat on the fatty acids of mature milk. A study of four ethnic groups. Am J Clin Nutr 1965; 17: 180–183. 3. Jensen R. G. The lipids in human milk. Prog Lipid Res 1996; 35: 53–92. 4. Schmeits B. L., Okolo S. N., VanderJagt D. J., Huang Y-S., Glew R. H. Selective retention of n-3 and n-6 fatty acids in human milk in the face of increasing proportions of medium chain-length

& 2001Harcourt Publishers Ltd

155

(C10–C14) fatty acids. Prostaglandins Leukot Essent Fatty Acids 1999; 61: 219–224. 5. Knox E., VanderJagt D. J., Shatima D., Huang Y-S., Chuang L-T. Glew R. H. Nutritional status and intermediate chain-length fatty acids influence on the conservation of essential fatty acids in the milk of northern Nigerian women. Prostaglandins Leukot Essent Fatty Acids 2000; 63: 195–202. 6. Schmeits B. L., Cook J. A., VanderJagt D. J., Magnussen M. A., Bhatt S. K., Bolik E. G., Huang Y-S. Fatty acid composition of the milk lipids of women in Nepal. Nutr Res 1999; 19: 1339–1348. 7. Jensen R. G. Patton S. The effect of maternal diets on the mean melting points of human milk fatty acids. Lipids 2000; 35: 1159–1161. 8. Folch J., Lees M. Sloane-Stanley G. H. A simple method for the isolation and purification of total lipids from animal tissues. J Biol Chem 1957; 226: 497–509. 9. Morrison W. R. Smith L. M. Preparation of fatty acid methyl esters and dimethylacetals from lipids with boron trifluoride methanol. J Lipid Res 1964; 5: 600–608. 10. Huang Y. S., Smith R. S., Redden P. R., Cantrill R. C., Horrobin D. F. Modification of liver fatty acid metabolism in mice by n-3 and n-6 D6 desaturase substrates and products. Biochim Biophys Acta 1991; 1082: 319–327. 11. Linder M. C. Nutrition and metabolism of fats. In: Linder M. C. ed. Nutritional Biochemistry and Metabolism. New York: Elsevier Science Publish Company, 1985; p. 35. 12. Bagdade J. D., Buchanan W. F., Pollare T., Lithell H. Abnormal lipoprotein phospholipid-composition in patients with essential hypertension. Atherosclerosis 1995; 177: 209–215. 13. Kulschar R., Engelmann B., Brautigam C., Duhm J., Thiery J., Richter W. O. Fast transmission of alterations in plasma phosphatidycholine/sphingomyelin ratio and lysophosphatidylcholine levels into changes of red-blood-cell membrane phospholipid-composition after low-densitylipopoprotien apheresis. Eur J Clin Invest 1995; 25: 258–265. 14. Holman R. T., Johnson S. B., Ogburg P. L. Deficiency of essential fatty acids and membrane fluidity during pregnancy and lactation. Proc Natl Acad Sci USA 1991, 88: 4835–4839. 15. Yazzie D., VanderJagt D. J., Isichei C., Glew R. H. Fatty acid composition of serum phospholipids of mother-infant pairs in northern Nigeria. Nutr Res 1996; 16: 1107–1118. 16. Enomoto T. M., Isichei C., VandeJagt D.J., Fry D. E., Glew R. H. Decreased polyunsaturated fatty acids in sickle cell anemia. J Trop Ped 1998; 44: 28–34. 17. Wijendran V., Bendel R. B., Couch S. C., Philipson E. H., Thomsen K., Zhang X. Maternal plasma phospholipid polyunsaturated fatty acids in pregnancy with and without gestational diabetes mellitus: relations with maternal factors. Am J Clin Nutr 1999; 70: 53–61. 18. Jensen R. G., Ferris A. M., Lammi-Keffe C. J., Henderson R. A. Lipids of bovine and human milks. J Dairy Sci 1990; 73: 223–240. 19. Glew R. H., Omene J. A., Vignetti S., D’Amico M., Evans R. W. Fatty acid composition of breast milk lipids of Nigerian women. Nutr Res 1995; 15: 477–489. 20. Prentice A., Jarjou L. M., Drury P. J., Dewit O., Crawford M. A. Breast milk fatty acids of rural Gambian mothers: effects of diet and maternal parity. J Pediar Gastroenterol Nutr 1989; 8: 486–490. 21. Van der Westhuyzen J., Chetty N., Atkinson P. M. Fatty acid composition of human milk from South African black mothers consuming a traditional maize diet. Eur J Clin Nutr 1988; 42: 213–220. 22. Kneebone G. M., Kneebone R., Gibson R. A. Fatty acid composition of breast milk from three racial groups from Penang, Malaysia. Am J Clin Nutr 1985; 41: 765–769.

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

156

Glew et al.

23. Sas M., Gellen J. J., Dustsin N., Tunkeynoon M., Chalapati S., Crawford M. A., Drury P. J., Lenihan T. Ayeni O., Pinol A. An investigation of the influence of steroidal contraceptives on milk lipid and fatty acids in Hungary and Thailand. Contraception 1986; 33: 159–173. 24. Chulei R., Xiaofang L., Hongsheng M., Xiuman M., Guizheng L., Gianhong D., DeFrancesco C. A., Connor W. E. Milk composition in women from five different regions of China: the great diversity of milk fatty acids. J Nutr 1995; 125: 2993–2998. 25. Genzel-Boroviczeny O., Wahle J., Koletzko B. Fatty acid composition of human milk during the first month after term and preterm delivery. Eur J Pediatr 1997; 156: 142–147.

26. Ratnayake W. M., Chen Z. Y. Trans, n-3 and n-6 fatty acids in Canadian human milk. Lipids 1996; 31: S279–S282. 27. Dutta-Roy A. K. Fatty acid transport and metabolism in the fetoplacental unit and the role of fatty acid-binding protein. J Nutr Biochem 1997; 8: 548–557. 28. VanderJagt D. J., Arndt C. D., Okolo S. N., Huang Y-S., Chuang L-T., Glew R. H. Fatty acid composition of the milk lipids of Fulani women and the serum phospholipids of their exclusively breastfed infants. Early Hum Dev 2000; 60: 73–87.

Prostaglandins, Leukotrienes and Essential FattyAcids (2001) 65(3), 147^156

& 2001Harcourt Publishers Ltd