Fatty Acids and Retinyl Esters of Rat Milk: Effectsof Diet and Duration of Lactation1

A. CATHARINE ROSS, MARIA E. DAVILA* ANDMARGOT P. CLEARY2*

Department of Physiology and Biochemistry, and Departmentof Pediatrics, The Medical College of Pennsylvania, Philadelphia, PA 19129 and 'Department of Nutrition and Food

Science, Drexel University, Philadelphia, PA 19104

ABSTRACT This study was initiated to explore the quantitative and qualitativedifferences in milk total fatty acids and milk retinyl esters when either hydrogenatedor nonhydrogenated fat is fed during pregnancy and lactation. Rats were fed dietscontaining 10% by weight of corn oil or partially hydrogenated corn oil. Milk wascollected on d 1, 8 and 14 of lactation and analyzed for protein, total fatty acids, fattyacid pattern, and retinyl ester pattern. Whereas diet produced no quantitative differences in milk protein or total fatty acids, the pattern of milk fatty acids varied significantly. Rats fed corn oil produced milk having more medium-chain saturated fattyacids, less long-chain monoenoic fatty acids, and more polyunsaturated fatty acidscompared to those fed hydrogenated corn oil. Rats fed hydrogenated corn oil produced milk fat having 21-26 % of the trans fatty acid, elaidic acid. Significant differences were also observed with duration of lactation: medium-chain fatty acids increased three to fourfold between d 1 and 8, where cts-monoenes and polyunsaturatedfatty acids declined. The pattern of milk retinyl esters strongly reflected, but was notidentical to, that of total milk fat. Comparing d 14 milk from rats fed corn oil withthat from rats fed hydrogenated corn oil, medium-chain esters of retinol constituted24 and 11% of total retinyl esters, whereas saturated long-chain fatty acid estersconstituted 52 and 44 %, respectively, trans Fatty acid esters of retinol comprised 24 %of vitamin A esters in milk of rats fed hydrogenated fat. These data provide evidencethat the composition of milk retinyl esters, as well as that of total milk fat, is determined both by the type of fatty acids from diet and from diet-related differences inde novo synthesis of fatty acids within the mammary gland and other tissues. J.Nutr. 115: 1488-1497, 1985.

INDEXING KEY WORDS hydrogenated fat • trans fatty acid

It is well documented that milk composition is highly variable among species andchanges during the course of lactation (1).In 1947, Parrish et al. (2) analyzed the vitamin A composition of bovine milk andreported that approximately 97% of thevitamin A is in the form of esterified retinol.The origin of these esters and the mechanism by which vitamin A is transferred fromblood to milk are not completely understood. Whereas newly assimilated dietary

vitamin A is present in plasma in the formof retinyl esters of long-chain fatty acidstransported by the chylomicron and itsremnant, vitamin A released from liver is in

©1985 American Institute of Nutrition. Received for publication21 December 1984. Accepted for publication: 31 July 1985.

'Supported by Grant HD-16484 from the National Institutes of Health.Dr. Ross is recipient of a Research Career Development Award,HD/AM-00691 from the National Institutes of Health.

"Present address: Hormel Institute, University of Minnesota, Austin,

MN 55912.

1488

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

FATTY ACID AND RETINYL ESTERS OF RAT MILK 1489

the form of retinol bound to its specifictransport protein, serum retinol-bindingprotein [RBP (3)]. It is this form thatcomprises about 95 % of vitamin A in plasma of fasting animals and in which vitaminA is thought to be delivered to most of itstarget organs. However, a smaller amount ofesterified retinol that is found in fastingplasma in association with lipoproteins orwith chylomicron remnants after mealscould also contribute to uptake by targetorgans. From experiments with the dairycow, Tomlinson et al. (4) suggested thatvitamin A is transferred to the lactatingmammary gland preferentially in the formof retinol. Vahlquist and Nilsson (5) havealso presented evidence obtained from thelactating rhesus monkey that, under normaldietary conditions, about 90 % of the vitamin A entering the lactating mammarygland is in the form of retinol. They alsoproposed that during periods of high vitamin A intake the proportion of vitamin Athat enters directly from lipoproteins, presumably the chylomicron, might be increased. Based on the assumption thatvitamin A is delivered to the lactating mammary gland either as retinol or as retinylesters that undergo hydrolysis during uptake, we postulated earlier that retinolesterification must take place in the mammary gland during lactation, and we demonstrated that membrane fractions prepared from mammary tissue of the lactatingrat are able to synthesize retinyl esters invitro (6). In this investigation, we also determined the composition of milk retinyl estersin a few samples of rat milk collected fromanimals fed a commercial nonpurified diet.We observed at least eight different fattyacid esters of retinol, including those with 8to 14 carbons, which accounted for nearlyone-fourth of the vitamin A esters in milk.Thus, the composition of the retinyl esters inrat milk differed markedly from the retinylester composition of intestinal chylomicronswhere only esters of long-chain fatty acids,principally palmitate, were found (7, 8).

It has been clearly shown that the composition of long-chain fatty acids of milkcan be changed significantly by varying thetype of fat in the maternal diet (9-11). Additionally, the presence of medium-chain fatty

acids in milk reflects the unique contribution of the mammary gland to milk synthesis. In the present study, we have sought tolearn more about the role of diet in determining the ester composition of total milkfat over the course of lactation and to obtainnew information about the composition ofmilk retinyl esters. We have chosen to compare two fats, corn oil having a high percentage of polyunsaturated fatty acids, andpartially hydrogenated corn oil in which thedegree of saturation but not of fatty acylchain length is modified. Because hydrogénation results in significant isomerization ofunsaturated fatty acids, we have been ableto ask whether the trans isomer of octa-decenoic acid is efficiently transferred tomilk and whether it, too, can be incorporated into milk retinyl esters. By collectingmilk from rats at d 1, 8 and 14 of lactation,we have also been able to obtain more detailed information about the changes inmilk fat composition that occur during thecourse of lactation in the presence of aconstant composition of dietary fat. Twolevels of dietary vitamin A have been employed to assess effects on the amount ofvitamin A in milk (12) and any effects ontotal fatty acids in milk.

MATERIALS AND METHODS

Design. The diets and animals used in thisstudy were part of an experiment to assessthe quantitative changes in milk vitamin Aand fat as well as qualitative changes duringthe course of lactation. A full description ofour experimental design, animals, diets,sources of materials and the effects of thesediets in the concentrations of vitamin A inmilk and on neonatal vitamin A status havebeen presented in reference 12. Briefly, 30adult female Sprague-Dawley rats (150-200 g)were purchased from Charles River Breeding Laboratories (Wilmington, MA) and, onarrival, were assigned to one of five dietarygroups. The four main diet groups constituted a 2 x 2 design in which we varied theamount of vitamin A [low vitamin A diethaving 0.6 retinol equivalents (R.E.) andhigh vitamin A diet having 15 R.E. per gramdiet in the form of retinyl acetate] and thetype of dietary fat (10% by weight of corn

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

1490 ROSS, DAVILA AND CLEARY

oil or of partially hydrogenated corn oil).The levels of vitamin A were chosen torepresent lower and higher intakes withinthe normal range rather than to be examplesof inadequate or excessive intake and, asshown (12), serum vitamin A concentrationsand growth were equivalent at both of theselevels of dietary vitamin A. In comparison,the National Research Council report hasrecommended 4 IU (1.2 R.E.)/g diet as anoptimal level for the rat (13). We determinedthe fatty acid compositions of the two fatsby gas-liquid chromatography (GLC) andfound for corn oil: palmitic acid, 12.0%;stearic acid, 1.8%; oleic acid, 25.8%; andlinoleic acid, 53.8 %; and for hydrogenatedfat: palmitic acid, 12.0%; stearic acid,18.0%; oleic acid, 28.3%; and elaidic acid(trans isomers of octadecenoic acid thatcomigrated with elaidic acid), 47.6%. Thefifth group of rats was included as an independent reference group primarily tocompare weight gains during pregnancyand lactation (12). Milk samples from thisgroup, when analyzed, were not included inthe statistical analysis of our 2x2 design.Diets were fed for 1 wk before breeding,throughout pregnancy and until d 14 of lactation. Milk samples were collected fromdams on d 1 (the day pups were first observed), d 8 and 14 of lactation using gentlehand-stripping of teats after stimulatingmilk let-down with oxytocin as previouslydescribed (12).

Determination of milk protein. The protein content of rat milk reported hereinwas determined by the method of Lowry etal. (14) as modified by Markwell et al. (15)to include sodium dodecylsulfate (NaDodSO4)as a solubilizing agent for lipid and protein.Bovine serum albumin was used as standard. Because of the high fat content of ratmilk, the unmodified procedure of Lowry etal. (14) without NaDodSO4 or the biuretprocedure (16) resulted in turbid reactionmixtures. Extraction of these samples withchloroform or diethyl ether either failedto completely remove turbidity or resultedin loss of color, whereas application ofthe NaDodSO4-Lowry procedure resulted inclearing of turbidity.

Determination of rat milk fatty acids.Total lipids were extracted from 50-100 ¡aof

rat milk by the method of Folch et al. (17)after addition of 1 mg of heptadecanoic acidas an internal standard. After separation ofsolvent phases and washing (17), trans-esterification was performed with the borontrifluoride-methanol reagent (Applied Science Laboratories, Inc., State College, PA)as described by Morrison and Smith (18),containing benzene to solubilize neutrallipid esters. Fatty acid methyl esters wereextracted into hexane, concentrated undernitrogen at 38-40 °Cand redissolved in

approximately 100 /il of hexane. Portions(1-2 ¡A)were separated by GLC on a 6-footx 1/8-inch glass column packed with acyanosilicone liquid phase (GP 3 %, SP-2310/2% SP-2300 on 100/200 Chromosorb W AWfrom Supelco, Inc., Belief onte, PA) withhelium as the carrier gas. The operatingconditions were as follows: injected anddetector temperatures were maintained at250 °C;the column temperature was 180°C

for the first 10 minutes of each analysisfollowed by a programmed temperatureincrease of 2.5°C/min to 220 °C,which washeld for 5 min before cooling and reequili-bration. This protocol allowed sufficienttime for elution of 22-carbon hexaenoicfatty acids in methyl ester standards. Peakareas were determined after flame ioniza-tion by an electronic integrator whose response was tested with quantitative methylester mixtures (NHI C from Supelco, Inc.;K-108 and H-104 from Applied Sciences,Inc.). For determination of the relativeamounts of cis and trans isomers of octadecenoic acid in the diet and milk of rats fedhydrogenated fat, portions of the samemethyl ester preparations were analyzed byGLC on a stainless-steel column of GP 15%OV-275 on 100/200 Chromosorb P AW-DMCS (Supelco, Inc.) using an initial column temperature of 230 °Cfor 25 min. The

percentages of oleic acid and trans octadecenoic acid (comigrating with elaidicacid) were calculated by measuring the relative areas under these peaks and multiplyingthese fractions by the percentage of totaloctadecenoic acid that was determined afterGLC on the GP 3% SP-2310/2% SP-2300column. Fatty acids have been assignedidentities on the basis of retention timesequal to those of authentic fatty acid methyl

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

FATTY ACID AND RETINYL ESTERS OF RAT MILK 1491

ester standards. The total fatty acid contentof milk has been expressed as the sum of allfatty acids having 10 or more carbons.

Analysis of rat milk retinyl esters. Thepattern of vitamin A esters was determinedby high performance liquid chromatogra-phy (HPLC) essentially as described by Ross(19). Two internal standards, retinyl hep-tanoate and retinyl pentadecanoate, wereadded to milk samples before extraction ofneutral lipids into hexane Retinyl esterswere then isolated as a group on columns ofaluminum oxide (6), concentrated undernitrogen, and dissolved in approximately200 /il of HPLC-grade acetonitrile Individual retinyl esters were separated by HPLCon a 5 IJLSupelcosil LC-8 column, 25 cmx 4.6 mm, and monitored by ultravioletabsorption at 340 nm. The mobile phasewas acetonitrile/water, 88:12 (vol/vol), untilretinyl oleate emerged, followed by thesame solvents in a 98:2 mixture to eluteretinyl stéarateNo longer chain esters ofretinol were found. Because the solubilitiesand, hence, recoveries of retinyl esters decreased slowly with decreasing polarity, wequantitated the medium-chain fatty acidesters of retinol (8 to 12 carbons) relativeto the recovery of the retinyl heptanoate,which they more closely resemble, and thelonger-chain esters, relative to retinylpentadecanoate

Statistical analysis. Data reported are themean ±SEM for each group. Differencesbetween diet groups and over time (tables 1,2 and 3) have been analyzed by a three-factor ANOVA where the factors were levelof dietary vitamin A, type of dietary fat,and either day of lactation in tables 1 and 3or type of fatty acid in table 2; this thirdfactor was treated as a repeated measuresfactor. Analyses were performed using theBiomédicalSeries P computer package P2V(20), followed by simple effects tests if interactions occurred. Differences in milk retinylester composition (table 4) were comparedbetween rats fed high vitamin A diets containing either corn oil or partially hydroge-nated corn oil because of the analyticalsensitivity of the HPLC method (19) andamount of sample available Data for eachretinyl ester were compared for significanceby using a two-way ANOVA with repeated

measures on one factor; in this model thetwo factors were type of fat in the diet andtype of retinyl ester in rat milk (10 types)where the percentage values of the variousretinyl esters were treated as 10 levels of arepeated measure factor. Simple effects testsand post-hoes (one-way ANOVA, Newman-Keuls test and independent t-test) wereperformed to follow up any statistically significant interactions of the main analysis ofvariance Details of analyses are presentedin the footnotes to tables.

RESULTS

Total milk fatty acid and protein concentrations. The concentration of fatty acidsand milk protein was determined in milkcollected on d 1, 8 and 14 of lactation (table1). We observed no significant differencesamong our four experimental diets in eitherof these constituents, as we expected onthese isocaloric diets. There also was nostatistically significant effect of day of lactation on these measurements although theaverage fatty acid concentration tended tobe greater on d 8 for each diet group.

Milk fatty acid composition. The composition of milk fat on d 1, 8 and 14 ispresented in table 2. All dams fed hydroge-nated corn oil had very different milk fattyacid patterns from dams fed corn oil. Thelevel of vitamin A did not affect total fattyacid patterns on d 1 and 8. On d 14 theamount of palmitic acid (16:0) was slightlygreater in milk of rats fed corn oil, whereasthat of stearic acid (18:0) was somewhatreduced; however, the level of significance ofthese effects was low (P < 0.05). In contrast,highly significant differences (P < 0.0001)were found for the interaction between typeof dietary fat and types of fatty acids inmilk. As shown in table 2, significant differences were observed not only for long-chainfatty acids derived exclusively or partly fromdiet, but also for medium-chain fatty acids,all of which had been synthesized withinmammary tissue. Compared to rats fed cornoil, those dams fed hydrogenated fat produced milk having a lower concentration ofmedium-chain fatty acids by d 8 and 14 oflactation, although differences were notfound in early rat milk (d 1). Rats fed

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

1492 ROSS, DAVILA AND CLEARY

TABLE 1Milk protein and total fatty add concentrations on d 1, 8 and 14 of lactation''1

Diet

Protein concn on Total fatty acid concn on

d 1 d8 d 14 d 1 d8 d 14

g/dl milk

High vitaminA,cornoilLow

vitaminA,cornoilHigh

vitaminA,hydrogenatedcorn

oilLowvitaminA,hydrogenatedcorn

oilNonpurified6.68.16.86.77.5±

0.3(5)±

1.5(5)±

0.4(5)±

0.8(5)±

1.5(2)7.3

±0.4(5)7.4

±0.2(5)7.6

±0.5(4)7.3

±0.5(5)8.0

±0.6(4)7.77.78.18.28.4±

0.3(5)±

0.1(5)±

0.4(6)±

0.2(6)±

0.5(5)17.5

±(5)16.6

±(5)14.1

±(6)16.5

±(5)10.6

±(4)0.91.41.01.81.919.6

±(5)21.8

±(4)17.4

±(4)18.4

±(5)15.9

±(4)1.52.62.11.81.715.4

±(5)18.4

±(5)15.7

±(6)14.0

±(6)12.1

±(4)1.31.71.91.31.4

'Results are means ±SEMfor the number of animals shown in parentheses. Statistical comparison of the

four groups fed purified diets indicated no significant differences among groups or over time. The group fed thestock diet was not included in this analysis of data.

hydrogenated corn oil also had a greaterpercentage of monounsaturated fatty acids(palmitoleic and oleic acids) on each day,and very much lower concentrations of thepolyunsaturated fatty acids. Additionally,the trans isomers of octadecenoic acid thatmigrated as methyl elaidate on GLC madeup 21-30% of total milk fatty acid throughout lactation for rats fed hydrogenated cornoil.

The composition of milk fat changedsignificantly with duration of lactation. Forstatistical analysis, we grouped milk fattyacids into five categories: medium-chainsaturated, long-chain saturated, cts-mono-enes, frons-monoenes and long-chain polyunsaturated fatty acids. Changes over timewere significant for three of these fatty acidgroups (table 3). First, the proportion (andabsolute amount) of medium-chain fattyacids increased very significantly betweend 1 and 8 of lactation. This change wasobserved in all diet groups but was mostpronounced among rats fed corn oil. By d 8,21-35% of total fatty acid mass was comprised of medium-chain fatty acids. Second,the proportion of cts-monoenoic fatty acidsgenerally decreased with duration of lactation. The third fatty acid group to showchanges with time of lactation was thepolyunsaturated fatty acid group, which

also decreased between d 1 and 8 of lactation. In contrast, total long-chain saturates(palmitic and stearic acids) and trans fattyacid from rats fed hydrogenated corn oilwere constant in proportion and in amountthroughout the 2 wk of observation.

Composition of retinyl esters in rat milk.We have analyzed the types and relativeamounts of vitamin A esters in milk todetermine whether ester composition can bemodified by diet and, if so, how retinyl esterpattern is related to the pattern of total fattyacids in milk. Figure 1 shows two typicalchromatograms of retinyl esters extractedfrom milk of rats fed corn oil or hydrogenated corn oil. Rat milk contains at leastnine different fatty acid esters of retinolranging in chain length from 8 to 18 carbonsand having 0-2 double bonds. Table 4 showsthat the quantitative distributions of retinylesters on d 8 and 14 of lactation do, indeed,vary significantly with the type of fat fed tothe lactating dam. Rats fed hydrogenatedfat had significantly lower amounts of retinyl decanoate (10:0), laurate (14:0) on d 14,and more retinyl oleate (18:1 cis). For damsfed hydrogenated corn oil, elaidic acid (18:1trans) was incorporated into milk retinylesters and constituted one-fourth of allvitamin A esters. In some samples, retinylelaidate was the major retinyl ester. We

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

FATTY ACID AND RETINYL ESTERS OF RAT MILK 1493

TABLE 2Fatty acid composition of rat milk on d 1, 8 and 14 of lactation'

Time of measureand fattyaciddl10:012:014:016:016:118.018:1

cis18:1trans18:2>18:23d810:012:014:016:016:118:018:1

cis18:1trans18:2>18:2dl410:012:014:016:016:118:018:1

cis18:1

trans18:2>18:2High

A,CO(5)2.58

±1.153.34±2.001.74±0.6422.32±1.553.60±0.402.94±0.1729.38±1.870.00±0.0026.80±0.866.80±2.29(5)8.48

±1.129.08±0.858.16±0.8122.78±2.231.54±0.282.98±0.2619.34±1.590.00±0.0024.32±1.622.88±1.47(5)9.28

±2.8211.32±0.4611.12±0.3520.34±0.921.36±0.132.08±0.4316.54±1.310.44±0.3122.64±1.623.72±2.10Fatty

acid as%Low

A,CO(5)4.36

±0.912.50±0.691.54±0.4222.22±2.283.90±0.543.16±0.1628.42±0.930.00±0.0026.08±1.576.76±1.47(5)12.92

±1.8811.90±0.469.96±0.7221.9±1.291.46±0.461.96±0.3517.00±1.304.26±4.2616.08±4.121.90±0.34(5)5.18

±0.799.14±1.0810.76±0.6223.84±1.472.06±0.442.82±0.1119.34±1.330.68±0.6825.14±0.711.04±0.32of

total, bydietHigh

A,HCO(6)2.15

±0.821.40±0.291.75±0.1222.15±1.506.73±0.295.70±0.6530.38±0.5924.47±2.222.42±0.611.53±0.61(4)10.38

±3.006.60±2.144.10±0.9119.60±1.883.68±0.523.58±1.0628.15±2.9522.68±2.360.58±0.580.18±0.18(4)3.83

±0.715.50±0.895.62±1.1720.83±0.953.92

±0.55.82±0.3826.63±1.9025.52±2.231.53±0.340.28±0.25Low

A,HCO(5)3.66

±1.432.22±0.602.00±0.1521.46±0.936.26±0.354.94±0.2530.04±0.6923.56±1.203.80±0.411.64±0.55(5)7.60

±1.26.00±1.175.00±0.7620.50±0.664.10±0.354.96±0.2927.10±1.6522.42±1.991.32±0.400.82±0.56(5)3.53

±1.097.72±1.328.97±1.3425.20±1.514.07±0.834.70±0.3421.10±2.8022.32±2.600.93±0.220.18±0.09Significance

ofCO vs. HCOS

PNSNSNSNS<0.0001<0.001NS<0.0001<0.0001<0.05NS=

0.01<0.001NS<0.001NS<0.01<0.0001<0.0001NS<0.05<0.05<0.05NS<0.01<0.001<0.01<0.001<0.0001<0.05

'Values are means ±SEMfor the number of animals shown in parentheses. Diet abbreviations: high A, highvitamin A; low A, low vitamin A; CO, corn oil; HCO, hydrogenated corn oil. 2On d 1 and 8, three-factor

analysis of variance was performed where the three factors were type of dietary fat, level of dietary vitamin A,and type of fatty acid in milk. The latter factor was treated as a repeated measures factor. On both d 1 and 8,significant effects were a main effect for milk fatty acid type (P < 0.0001) indicating differences in the levelsof individual fatty acids in milk and an interaction between dietary fat and milk fatty acid type (P < 0.0001).These interactions were followed with a one-way analysis of variance on all four diet groups (to maximize power)for each milk fatty acid followed by planned contrasts to compare rats fed corn oil vs. hydrogenated corn oil.The significance of this comparison is reported above. NS, not significant. Differences in the fatty acid composition of rat milk on d 14 were tested for significance using the statistical analysis described above. The results ofthe main analysis of variance revealed a main effect of the type of fatty acid in milk (P < 0.0001), an interactionbetween fatty acid type in milk and type of dietary fat (P < 0.0001), a weaker interaction (P < 0.05) betweentype of dietary fat and level of vitamin A, and an interaction (P < 0.02) between all three factors. We followedup with a two-factor (type of fat, level of vitamin A) analysis of variance within each milk fatty acid to determinefor which of the milk fatty acids the four experimental groups differed. The results of this test for the comparisonof rats fed corn oil vs. hydrogenated corn oil are reported above. On d 14 only, there was also a main effect forthe amount of vitamin A in the maternal diet on the percentage of palmitic and stearic acid in rat milk. At present, it is not known why the level of vitamin A had an effect on these two milk fatty acids on this day only.The effect was significant only at P < 0.05 and may be spurious. 3The sum of all fatty acids having retentiontimes longer than that of methyl linoleate by gas-liquid chromatography.

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

1494 ROSS, CAVILA AND CLEARY

TABLE 3Changes with day of lactation in the fatty acid composition of rat milk1''

Fatty acidgroupMedium-chainsaturated(10:0,

12:0,14:0)Long-chainsaturated(16:0,

18:0)cis-Monoenoic(16:1,

18:1)trani-Monoenoic(elaidic

acid)Long-chainpolyunsaturated(18:2,

>18:2)Fatty

acid groupasDietHigh

A, CO(5)4Low

A, CO(5)HighA, HCO(4)LowA, HCO(4)High

A,COLowA,COHighA,HCOLowA,HCOHigh

A,COLowA,COHighA,HCOLowA,HCOHigh

A,COLowA,COHighA,HCOLowA,HCOHigh

A,COLowA,COHighA,HCOLowA, HCOdl7.66

±2.50"8.40±1.56"5.20±1.36'7.35±2.60*25.26

±1.6325.38±2.2128.85±2.1625.90±1.1232.98

±2.16'32.32±1.17*37.55

±0.8636.53±1.04'0.00

±0.000.00±0.0023.08±3.2323.65±1.5433.60

±3.0'32.84±2.28"3.43

±1.516.10±0.84"25.7234.7821.0821.4025.7623.9023.1825.6820.8818.4631.8330.030.004.2622.6820.7827.2017.980.751.95%

oftotald8±

2.41b±2.14b±5.33b±3.38b±

2.46±1.41±2.28±0.55±

1.80b±1.76b±

3.41±1.87b±

0.00±4.26±2.36±1.45±

2.47b±4.02b±

0.75±0.95bfatty

acidson41431.72

±3.03b25.08±2.05°14.08±3.66""19.43±4.10"22.42

±1.1126.66±1.5327.30±0.3028.93±1.4917.90

±1.40b21.40±1.64b32.30

±3.1523.58±2.23b0.44

±0.310.68±0.6823.60±2.9625.28±2.8526.36

±1.43"26.18±0.46e2.15

±0.241.35±0.23bSignificance

of time oflactation3P<0.0001<0.0001<0.02<0.02NSNSNSNS<0.001<0.001NS<0.001NSNSNSNS<0.05<0.001NS<0.01

'Values are means ±SEM.CO, corn oil diet; HCO, hydrogenated corn oil diet. 'Groups in the same rowbearing different superscripts are significantly different, P < 0.05, by the Newman-Keuls test. 'Results of simpleeffect across time for each cell of the design. 'Only animals for which all 3 d of milk collection were available

for analysis are included in this analysis of data.

observed only two polyunsaturated esters ofretinol, retinyl linoleate (18:2) and, in someanalyses, traces of retinyl linolenate (18:3).The esterification of retinol with these polyunsaturated fatty acids of dietary origin wassignificantly greater in milk of rats fed thecorn oil diet. On either type of dietary fat,nearly all retinol in milk was present inesterified form, averaging 97 % for five ratsfed hydrogenated corn oil and 110% for fiverats fed corn oil.

DISCUSSION

The fatty acid composition of lipids innumerous mammalian tissues, membranefractions, depot fat, or secretions such aslipoproteins and milk has been analyzedand, often, composition has been found tobe affected by the type of dietary fat. Yet,

little is understood of the biochemical basisfor fatty acid specificity or the manner inwhich diet modifies this mechanism (s). Thecomposition of milk fat has been shownpreviously to vary with caloric balance (9)and with the type (9-11, 21-24) and amountof fat in the maternal diet (9, 10). We beganthis study to examine temporal change inmilk fat composition in the rat and its relationship to diet, and to determine milkretinyl ester composition on diets containingdifferent fats. In our study, diets were fed toall animals for approximately four weeksprior to the lactation period and thus represent chronic dietary regimens. The twolevels of vitamin A were chosen to representintakes in the upper and lower ends of anormal range rather than to produce vitamin A toxicity or deficiency and, indeed, wehave shown that both growth rates and

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

FATTY ACID AND RETINYL ESTERS OF RAT MILK 1495

EcO

m

SO 005 rr Mllk "'t'nyl Ester PatternRat Ftd HydrooenotedCorn Oil

10:080

M\l

180

50

Volume, ml

Fig. 1 Chromatogram of milk retinyl esters fromrats fed corn oil (1A) and hydrogenated corn oil (IB) asdietary fat. Internal standards of retinyl heptanoate(7:0) and retinyl pentadecanoate (15:0) are shown in 1Ain dashed lines. Retinyl oleate is shown by 18:1 c andretinyl elaidate by 18:1 t. Arrow denotes change ofsolvent composition (see the materials and methodssection), and 50% decrease of sensitivity.

plasma vitamin A levels were equal andnormal at both levels of vitamin A (12). Thelevels we employed did not have any significant effects on the amount of milk protein or milk fat, or on the pattern of totalmilk fatty acids. We also found no significant quantitative changes with duration oflactation in either the concentration of milkprotein or milk fatty acids, although therewas a trend toward a higher protein concentration later in lactation in most of our dietgroups (table 1) similar to the observation ofKeen et al. (25), and toward a greater fat

concentration on d 8 of lactation. One mayinfer from these values that the rat does notsecrete a true protein-rich colostrum or doesso for only a very short time.

Despite the constancy of total fat concentration, an examination of fatty acid patternindicates that there are striking developmental changes in lipid secretion duringlactation. To analy/e these changes, we havegrouped milk fatty acids into five categories(table 3) that comprise similar physical

TABLE 4Distribution of vitamin A esters in rat milk1

Time ofmeasure

and retinylestertyped8

8:010:012:014:016:016:118:018:1

cis18:1iron«18:2d

148:010:012:014:016:016:118:018:1

c«18:1troni18:2Retinyl

ester as % of totalby dietaryfatCO(4)

3.01 ±0.667.69±1.112.78±0.575.47±0.5033.49±1.832.05±0.7820.83±1.6911.99±1.470.00±0.0012.35±1.49(5)

4.13 ±0.6410.08±1.454.20±0.815.28±0.7237.65±1.311.47±0.2914.20±1.248.3±0.620.0±0.0013.9±0.85HCO(4)

1.88 ±0.432.34±0.531.61±0.472.99±0.4728.33±3.292.25±0.3711.87±1.4620.32±1.4125.62±1.622.61±0.52(6)

2.31 ±0.523.82±0.521.68±0.223.52±0.4631.19±2.882.07±0.3213.04±2.3816.9±1.9824.1±2.162.1±0.55Signifi

cance oftype ofdietaryfat2PNS<0.01NS<0.01NSNS<0.01<0.01<0.0001<0.001NS<0.01<0.01NSNSNSNS<0.01<0.0001<0.001

'Because of the analytical sensitivity of the high-

performance liquid chromatography method (19), amountof milk available, and our finding that the level of vitaminA had no effect (d 8) or a very slight effect (d 14) on totalfatty acid patterns in milk, only milk from rats fed the higherlevel of vitamin A with either corn oil (CO) or hydrogenatedcorn oil (HCO) was analyzed and included here. Values aremeans ±SEMfor the number of rats stated in parentheses.2A two-factor (type of dietary fat and type of retinyl ester)

analysis of variance was performed for each day separately; on each day, there was a strongly significant interaction,P < 0.0001. These interactions were followed up with unpaired two-tailed i-tests for each retinyl ester and theseresults are presented above. NS, not significant.

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

1496 ROSS, DAVILA AND CLEARY

characteristics or metabolic origins. In thecase of medium-chain fatty acids, all werederived from de novo synthesis in the breast,for our dietary fats had only 16- and18-carbon fatty acids. Both elaidic acid andlinoleic acid originate from diet exclusively,either by direct uptake from the chylo-micron or from plasma lipoprotein afterreincorporation in the liver. In the case ofother fatty acids (long-chain saturates andcis-monoenes) it is not possible from thesemass measurements to distinguish betweenan origin in the mammary gland, synthesisin other tissues, or assimilation of newlyabsorbed dietary fat. For those fatty acidsknown to be of either mammary glandorigin or dietary origin, we observed twovery different temporal patterns. Betweend 1 and 8 of lactation, we observed a 3.7-fold increase in total medium-chain fattyacids in milk of rats fed corn oil (table 3),with an especially notable change in myris-tic acid (table 2). We did not measure milkyield in this study, but data of Chalk andBailey (26) indicate a 1.5- to 2-fold increaseduring this time and, thus, total secretion ofmedium-chain fatty acids increased nearly6-fold during the first week of lactation. Incontrast, the concentration of trans isomersof octadecenoic acid in milk of those ratsfed partially hydrogenated fat was highlystable, equaling 21-26% of total fatty acidson each day of lactation. Linoleic aciddecreased somewhat as lactation progressedin all groups. These data indicate thatuptake of fatty acids of dietary origin occursessentially in proportion to milk yield, ordecreases very slightly, throughout the lactation period.

An additional observation is that the concentrations of medium-chain fatty acidswere significantly lower in rats fed hydrogenated fat, although the developmentalchange toward higher concentrations wasstill observed (tables 2, 3). It is possible thatthis interaction of diet with fatty acidsynthesis is mediated by one or more of theproducts of hydrogénation, for Grigor andWarren (11) have also observed a decreasedrate of lipogenesis in mammary tissue fromlactating rats fed hydrogenated coconut oil.

Our analysis of retinyl ester pattern ismost notable for two features. First, several

esters were observed in milk that have notbeen reported previously in other biologicalsamples, namely medium-chain fatty acidesters and retinyl elaidate. This variety inretinyl ester composition is in marked contrast to results reported by Goodman andco-workers (7, 8), who observed only long-chain fatty acid esters of retinol in lymphchylomicra with retinyl palmitate predominant, and of Bridges and others (27) whohave found, almost exclusively, retinyl estersof palmitate and stéaratein the eye. Others(28) have shown a preference for retinyl palmitate in liver vitamin A stores. Secondly,diet had a marked effect on the retinyl esterpattern observed in milk. This is, again, incontrast to lymph chylomicra where thetype of fat in a test meal had little effect onretinyl ester pattern (8). It is somewhatsurprising to us that the milk fat globule andchylomicra, both of which are triglyceride-rich lipoproteins secreted for fat transportand in which triacylglycerol fatty acidcomposition reflects diet should have retinylester compositions that are, in one case(milk) diet dependent and, in the othercase, quite diet independent. The mechanisms responsible for these fatty acid specificities are not yet known. It is of possibleinterest that Ong (29) has reported recentlythe presence of a retinol-binding protein inthe intestine that may participate in absorption or the esterification reaction. Wehave reported previously the presence of anesterifying enzyme (acyl-CoA:retinol acyl-transferase, ARAT) in mammary gland (6)and liver (30) membranes, and others havereported similar findings for the intestine(31). Whether retinyl ester type is determined by this or another acyltransferase,whether specific lipid-binding proteinsdetermine retinyl ester type, and whethermechanisms for retinol esterification are essentially the same or different in various tissues remain interesting questions for futureinvestigations.

ACKNOWLEDGMENT

We thank Mr. Ed Gracely of The MedicalCollege of Pennsylvania's Biostatistics Center

for his helpful advice

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from

FATTY ACID AND RETINYL ESTERS OF RAT MILK 1497

LITERATURE CITED

1. Jenness, R. (1974) The composition of milk. In:Lactation: A Comprehensive Treatise, vol. 3,(Larson, B. L. & Smith, V. R., eds.), pp. 3-107,Academic Press, New York.

2. Parrish, D. B., Wise, G. H. & Hughes, J. S.(1947) The state of vitamin A in colostrum andin milk. J. Biol. Chem. 167, 673-678.

3. Goodman, D. S. (1984) Plasma retinol-bindingprotein. In: The Retinoids, vol. 2, (Sporn, M. B.,Roberts, A. B. & Goodman, D. S., eds.), pp. 41-88,Academic Press, New York.

4. Tomlinson, J. E., Mitchell, G. E., Jr., Bradley,N. W., Tucker, R. E., Boling, J. A. & Schelling,G. T. (1974) Transfer of vitamin A from bovineliver to milk. J. Anim. Sci. 39, 813-817.

5. Vahlquist, A. & Nilsson, S. (1979) Mechanismsfor vitamin A transfer from blood to milk in therhesus monkey. J. Nutr. 109, 1456-1463.

6. Ross, A. C. (1982) Retinol esterification bymammary gland microsomes from the lactatingrat. J. Lipid Res. 23, 133-144.

7. Huang, H. S. & Goodman, D. S. (1965) Vitamin A and carotenoids. I. Intestinal absorptionand metabolism of "C-labeled vitamin A and ß-carotene in the rat. J. Biol. Chem. 240, 2839-2844.

8. Goodman, D. S., Blomstrand, R., Werner, B.,Huang, H. S. & Shiratori, T. (1966) The intestinal absorption and metabolism of vitamin A and/3-carotene in man. J. Clin. Invest. 45, 1615-1623.

9. Insull, W, Jr., Hirsch, T., James, T. & Ahrens,E. H., Jr. (1959) The fatty acids of humanmilk. II. Alterations produced by manipulation ofcaloric balance and exchange of dietary fats. J.Clin. Invest. 38, 443-450.

10. Brandorff, N. P. (1980) The effect of dietary faton the fatty acid composition of lipid secreted inrat's milk. Lipids 15, 276-278.

11. Grigor, M. R. & Warren, S. M. (1980) Dietaryregulation of lipogenesis in lactating rats. Bio-chem. J. 188, 61-65.

12. Davila, M. E., Norris, L., Cleary, M. P. & Ross,A. C. (1985) Vitamin A during lactation: relationship of maternal diet to milk vitamin Acontent and to the vitamin A status of lactatingrats and their pups. J. Nutr. 115, 1033-1041.

13. National Research Council (1978) Nutrientrequirements of the laboratory rat, no. 10, pp.7-37, National Academy of Sciences, Washington,DC.

14. Lowry, O. H., Rosebrough, N. J., Farr, A. L. &Randall, R. J. (1951) Protein measurementwith the Folin phenol reagent. J. Biol. Chem. ¿93,265-275.

15. Markwell, M. A. K., Haas, S. M., Bieber, L. L. &Tolbert, N. E. (1978) A modification of theLowry method to simplify protein determinationin membrane fractions and lipoproteins. Anal.Biochem. 87, 206-210.

16. Cornali, A. G., Bardawill, C. J. & David, M. M.(1949) Determination of serum proteins bymeans of the biuret reaction. J. Biol. Chem. ¿77,751-766.

17. Folch, J., Lees, M. & Sloane Stanley, G. H.(1957) A simple method for the isolation andpurification of total lipides from animal tissues. J.Biol. Chem. 226, 497-509.

18. Morrison, W. R. & Smith, L. M. (1964) Preparation of fatty acid methyl esters and dimethyl-acetals from lipids with boron trifluoride metha-nol. J. Lipid Res. 5, 600-608.

19. Ross, A. C. (1981) Separation of long-chainfatty acids of retinol by high-performance liquidchromatography. Anal. Biochem. 115, 324-330.

20. Brown, M. B. (1977) BMDP-77: BiomédicalComputer Programs P Series, University of California Press, Berkeley, CA.

21. Mellies, M. J., Ishikawa, T. T., Gartside, P. S.,Burton, K., MacGee, J., Allen, K., Steiner, P. M.,Brady, D. & Gleuck, C. J. (1979) Effects ofvarying maternal dietary fatty acids in lactatingwomen and their infants. Am. J. Clin. Nutr. 32,299-303.

22. Sanders, T. A. B., Ellis, F. R. & Dickerson, J. W. T.(1978) Studies of vegans: the fatty acid composition of choline phosphoglycerides, erythrocytes,adipose tissue, and breast milk, and some indicators of susceptibility to ischemie heart disease invegans and omnivore controls. Am. J. Clin. Nutr.31, 805-813.

23. Green, M. H., Dohner, E. L. & Green, J. B.(1981) Influence of dietary fat and cholesterol onmilk lipids and on cholesterol metabolism in therat. J. Nutr. Ill, 276-286.

24. Craig-Schmidt, M. C., Weete, J. D., Faircloth,S. A., Wickwire, M. A. & Livant, E. J. (1984)The effect of hydrogenated fat in the diet ofnursing mothers on lipid composition and prosta-glandin content of human milk. Am. J. Clin. Nutr.39, 778-786.

25. Keen, C. L., Lönnerdal, B., Clegg, M. & Hurley,L. S. (1980) Developmental changes in composition of rat milk: trace elements, minerals, protein, carbohydrate and fat. J. Nutr. Ill, 226-230.

26. Chalk, P. A. & Bailey, E. (1979) Changes in theyield, and carbohydrate, lipid and protein contentof milk during lactation in the rat. J. Dev. Physiol.1, 61-79.

27. Bridges, C. D. B., Alvarez, R. A. & Fong, S.-L.(1982) Vitamin A in human eyes: amount, distribution, and composition. Invest. Ophthalmol.Visual Sci. 22, 706-714.

28. Futterman, S. & Andrews, J. S. (1964) Thecomposition of liver vitamin A ester and the synthesis of vitamin A esters by liver microsomes. J.Biol. Chem. 239, 4077-4080.

29. Ong, D. E. (1984) A novel retinol-binding protein from rat: isolation and partial characterization. J. Biol. Chem. 259, 1476-1482.

30. Ross, A. C. (1982) Retinol esterification by ratliver microsomes. Evidence for a fatty acyl-CoA:retinol acyltransferase. J. Biol. Chem. 257, 2453-2459.

31. Helgerud, P., Peterson, L. B. & Norum, K. R.(1983) Retinol esterification by microsomes fromthe mucosa of human small intestina J. Clin.Invest. 71, 747-753.

by guest on February 24, 2014

jn.nutrition.orgD

ownloaded from