Low-birth-weight (LBW) infants have high energy requirements and are dependent on high fat intake to maintain adequate postnatal growth. Fat energy is transported in plasma as triglycerides, which are either derived from the diet or from de novo lipogenesis (DNL). Hepatic DNL is a quantitatively minor pathway in healthy adults on a relatively high-fat Western diet when the energy intake is not in excess of total energy expenditure. However, hepatic DNL increases during critical illness in the absence of excessive caloric intake.1,2 Additionally, in animal and human studies, DNL increases with high-carbohydrate, low-fat diet or when the total caloric intake is in excess of energy expenditure.3-5 Thus, DNL may be influenced by both dietary fat composition and energy intake among other factors.
The actual contribution to plasma fatty acids from diet and DNL during breast feeding or receiving parenteral nutrition (PN) is unknown. Measurement of DNL in stable LBW infants with an adequate intake of calories from fat in breast milk or PN may provide insight into the role of DNL during normal postnatal growth. Fractional hepatic DNL can now be measured by a novel and noninvasive isotopic approach using mass isotopomer distribution analysis.6-8 We hypothesize that DNL plays an important physiologic role in adapting to exclusive breast milk (EBM) feeding or to PN. We measured fractional hepatic DNL and the contribution of endogenous substrate to acetyl CoA in DNL. The study subjects were stable LBW infants at similar postnatal age and were receiving nutritional energy intake of protein, carbohydrate, and fat by either breast-milk feedings or IV fat emulsion with PN.
MATERIALS AND METHODS
Subjects
The study was approved by the UCLA institutional review board of the office for protection of research subjects. Informed consent was obtained from parents after explaining study details. LBW appropriate for gestational age infants born at <34 weeks of gestation and receiving EBM feedings or completely receiving PN support were studied from the neonatal intensive care units at the Mattel Children's Hospital and Santa Monica UCLA hospital. All subjects were in stable cardiopulmonary status without need for ionotropic support or assisted mechanical ventilation. The nutrient and energy intake from protein, carbohydrate, and fat was unchanged for 4 days before the study. There was no evidence of postnatal growth failure (or weight less than fifth percentile).9 Infants with intrauterine growth failure, congenital heart disease, evidence of hepatic dysfunction, and metabolic disorders were excluded from the study.
Experimental Protocol
The experimental protocol is outlined in Figure 1. Infants were receiving a stable daily dietary intake of fat calculated as grams per kilogram for 4 days before the study. The average of daily fat intake was retrospectively calculated from the records. Stable isotope tracer [2-^sup 13^C] acetate (Cambridge Isotope Laboratories, Andover, MA; >98% pure), was administered orally or by IV. The dose was calculated to achieve a 10% enrichment of 2 carbon units of daily fat intake. The isotopic enrichment was continued for 72 hours. All oral and IV isotope solutions were prepared and tested for pyrogenicity and sterility before use by the UCLA research pharmacy. For the oral administration, the total daily dose was divided into 8 individual doses and dispensed in eight 1-mL syringes by the pharmacist. The neonatal intensive care unit (NICU) nurse gave 1 dose with each EBM feeding every 3 hours via the NG tube. The NG tube was then flushed with 2-3 mL of formula or saline to ensure administration of a full dose. For IV administration, the total daily dose was given by a continuous infusion via the preexisting central line for PN and IV lipids (IL) for 72 hours. Blood samples (100 µL of plasma) were collected at baseline or before tracer administration and then at 12, 24, 48, and 72 hours after beginning tracer administration by heel stick in the absence of a central line or via preexisting lines for blood drawing. The plasma was separated and stored at -4°C until analyzed.
Gas Chromatography-Mass Spectrometry (GC-MS) Analysis
Lipid extraction was performed using methods described by Lowenstein et al.10 GC-MS analysis was performed on a Hewlett-Packard model 5973 Mass Selective Detector connected to a model 6890 gas chromatograph (Palo Alto, CA) using electron impact ionization. A capillary column BPX70 (SGE, Austin, TX) measuring 30 m × 250 µm (ID) was used to separate fatty acid methyl esters. The GC conditions were carrier gas (helium) flow rate, 1 mL/min; injector temperature, 250°C; and oven temperature, programmed from 120-220°C at 5°C/min. Cholesterol was separated without a derivative on a SAC5 glass capillary column measuring 15 m × 0.25 mm × 0.25 µm film thickness (Supelco). The GC conditions were carrier gas (helium) flow rate of 1 mL/min; injector temperature, 270°C; and oven temperature programmed from 170-278°C at 35°C/min ramp and then hold for 6 minutes. After determining mass isotopomer distribution from the respective mass spectra,8,11,12 we calculated the average number of ^sup 13^C incorporated per molecule, and the fractional new synthesis.8,11-13
Analysis of Fatty Acid Composition
The total ion chromatogram of fatty acids in plasma allows the quantification of relative amounts of each fatty acid. The major peaks in the ion chromatogram are the long-chain fatty acids palmitate (C16:0) and stearate (C18:0). Other long-chain and polyunsaturated fatty acids were also detected in small amounts, representing <10% of the total fatty acids. The area under each peak was integrated using the ChemStation software. The long-chain and monounsaturated fatty acid distribution was calculated as a percent relative to the palmitate value.10,11,12
Statistics
All serum fatty acid measurements were made in triplicates. Coefficient of variation for mass isotopomer having enrichment over 1% was 10% or less for GC-MS analysis in our laboratory. All statistical data from EBM and PN groups for clinical characteristics and dietary intake are shown as mean ± SD, and ANOVA models were used followed by the Fisher PLSD test to validate intergroup comparisons. The percent fatty acid breakdown and fresh new synthesis are expressed as mean ± SEM. All comparisons of fresh new synthesis between groups were done with 2-sample Student's t tests. p Values were considered significant if <0.5.
RESULTS
The birth weight, gestational age, and brief clinical data for 8 LBW infants receiving exclusive EBM feedings (EBM group) fortified with human milk fortifier (Ross Premature formula products) to yield 24 calories per ounce and 6 LBW infants receiving exclusive PN and IL support (PN group) are shown in Table I. All infants in EBM and PN groups were born at <34 weeks of gestation, and the average birth weight in the EBM group was significantly lower than the PN group. However, at the time of study infants in both groups were of same postnatal age (34 ± 1.4 weeks and 34.2 ± 2.3 weeks, respectively). None of the study infants showed evidence of postnatal growth failure (body weight less than fifth percentile).9 Before the study, the PN infants required significantly more days of mechanical ventilation. There was no difference in supplemental oxygen and total days of PN support. All infants in both groups were weaned from mechanical ventilation before study. Five of 6 LBW infants in the PN group underwent closure of gastroschisis >5 days before this study, and 1 infant had repair of ileal atresia. The supplemental oxygen was weaned in all except for 2 LBW infants in EBM group.
The contributions of protein, carbohydrate, and fat energy intake as percent of total nutritional energy intake in the EBM and PN groups are shown in Table II. These remained unchanged for 4 days before the study and throughout the 3 days of isotope administration. The total weight gained during these 7 days of nutritional monitoring was used to calculate the average weight gain. The average weight gain was higher in the EBM infants compared with the PN group (36 ± 3.6 g vs 9.2 ± 3.6 g, p < .001) suggesting that the PN group may have higher energy demand at the time of study or the metabolic efficiency of PN is low because both groups were receiving similar total calories (Table II). All EBM feedings were administered by bolus orogastric gavage tube every 3 hours. The PN infants received 20% lipid emulsion (Baxter) and were receiving full calories from PN and IL. There was no difference in total protein intake between the 2 groups. The EBM group received a significantly higher fraction of calories from fats (49.3% ± 2.1 and 28.6% ± 1.5 respectively, p < .0001) and lower percent calories from carbohydrate (p < .001) compared with the PN group (Table II). The plasma triglyceride levels were within normal limits and in the range of 34-126 mg/dL in the breast-milk and PN groups. However, the plasma triglyceride levels were significantly higher in the EBM group compared with the PN group (Table II, p < .001).
Fatty acid composition of serum (Table III) shows the distribution of fatty acids (in percent) in infants receiving EBM feedings and receiving PN. The IL emulsions have a significantly larger fraction of C18:2 and other polyunsaturated fatty acids compared with EBM,14 yet the plasma fatty acid distributions within the 2 groups are very similar, suggesting adjustment of plasma lipid fatty acid composition by the liver.
The intake of [2-^sup 13^C] acetate introduced ^sup 13^C-labeled acetyl-CoA to the precursor pool. Precursor enrichment was calculated from the consecutive mass isotopomer ratio in palmitate. This steady-state molar enrichment is a function of the mixing of [2-^sup 13^C] acetate and unlabeled acetyl-CoA generated from glucose or fatty acids from the diet. The time course of ^sup 13^C acetyl-CoA enrichment in palmitate of LBW infants is shown in Figure 2. The plasma fatty acid enrichment achieved a steady state after 12 hours of isotope administration and remained relatively unchanged for the 72 hours of the experiment. EBM infants show relatively constant fractional enrichment at 10% ± 1.0% by [2-^sup 13^C] acetate and the PN infants at 7.5% ± 1.0%. The fractional enrichment in PN infants was lower (p < .007). However, 3 of 6 PN infants required diuretics during the study period. The average enrichment on infants not requiring diuretics was similar in 2 groups. Because the dose of [2-^sup 13^C] acetate was calculated according to intake to achieve a target enrichment of 10% for all infants, the lower enrichment in the PN group may be caused by urinary loss of [2-^sup 13^C] acetate.
The results of [2-^sup 13^C] acetate incorporation into palmitate and stearate are shown in Figures 3A and B. The de novo synthesis of palmitate and stearate increased steadily throughout the study period, whereas steady-state enrichment was attained in 12 hours. Palmitate and stearate synthesis in EBM infants at 72 hours (13.6 ± 2.5% and 11.2 ± 2.7%, respectively) was similar to the PN group (14.9% ± 0.6% and 10.6% ± 1.4%, respectively). Despite the similarity in DNL, infants in the EBM group maintained a much higher level of plasma triglycerides compared with infants in the PN group. In humans, DNL is known to increase in subjects receiving diets with high carbohydrate to fat ratio.3-5 The lack of difference in DNL between the EBM group receiving a high-fat (49.3% ± 2.1%) energy diet and the PN group receiving a low-fat (28.6% ± 1.5%) energy diet suggests that there is lack of compensatory change in DNL in the PN-group infants. The inability to compensate for a lower fat intake in the PN group may also explain the lower triglyceride level in these infants.
The new synthesis of stearate by chain elongation from available palmitate was determined from the ratio of the consecutive isotopomer ratio m^sub 2^/m^sub 1^ of stearate against m^sub 2^/m^sub 1^ of palmitate.13,15 In DNL, m^sub 2^/m^sub 1^ ratio is a function of precursor enrichment, and there is a constant relationship between m^sub 2^/m^sub 1^ from palmitate to m^sub 2^/m^sub 1^ ratio from stearate represented by the theoretical straight line in Figure 4A and B. Synthesis of stearate by chain elongation introduces only m^sub 1^ but not m^sub 2^ isotopomer, resulting in a lower m^sub 2^/m^sub 1^ ratio. In the EBM group, chain elongation is the more prevalent form of stearate synthesis, as represented in Figure 4A, because the m^sub 2^/m^sub 1^ ratio remains below the theoretical relationship for 100% DNL. In the PN group (Fig. 4B) the m^sub 2^/m^sub 1^ ratio remains similar to the predicted theoretical relationship for 100% DNL representing new synthesis. The availability of long-chain fatty acids in breast milk allows the EBM group to have more stearate synthesis through chain elongation than the PN group.
The average fractional contribution of newly synthesized cholesterol by [2-^sup 13^C] acetate incorporation was 12.7% ± 2.1% in all EBM infants and 17.4% ± 4.6% in PN infants (Fig. 5). Therefore, rate of cholesterol synthesis is similar in the 2 groups, but EBM infants have a higher dietary serum cholesterol intake compared with PN.
DISCUSSION
In animal studies, DNL is known to increase when fat energy in the diet is insufficient to meet the requirement for peripheral tissues. Because it is an energy-requiring process, it is suppressed by the administration of an isocaloric high-fat diet.13 Such an observation supports the concept that triglycerides are important energy-dense substrates transported in plasma. In healthy adults receiving a Western diet, the baseline DNL levels for palmitate are lower than 5%.1,8 However DNL is stimulated with the intake of a high carbohydrate (55%-75%), and low-fat (10%) diet3-5 when dietary fatty acids available for transport are relatively low. DNL is also increased in critically ill adult patients when basal energy expenditure is high.2 DNL provides the caloric dense substrate for transports to meet their increased energy requirement for the same cardiac output.
In the present study using mass isotopomer distribution analysis and stable isotope [2-^sup 13^C] acetate enrichment, we evaluated the hepatic DNL in stable LBW infants receiving isocaloric diet with different fat-energy intake. The infants received standard recommended intake of dietary fat and demonstrated normal postnatal growth. Under steady-state conditions of calorie and fat intake, the continuous administration of [2-^sup 13^C] acetate increased precursor enrichment over the first 12 hours and remained at steady state during the 3 days of study. About 10% of the acetyl-CoA contributing to plasma palmitate was derived from [2-^sup 13^C] acetate. However, DNL continued to increase after 12 hours of steady-state enrichment to increase to levels of 10-25% during first 24 hours. Our study shows a lack of compensatory down-regulation of DNL with dietary fat-energy intake of 30% to 60%, suggesting a very high percent fat-energy demand in both EBM and PN infants in order to sustain postnatal growth to match in utero fetal growth rate.
The LBW infants in the present study received either the natural diet of breast milk (49.3% ± 2.1% of fat calories) or standard recommended intake of lipids16 from the PN and IL (28.6% of fat calories or 3.55 g/kg/d of IL). Assuming that the total energy delivered by plasma triglycerides is represented by dietary energy intake plus DNL, it can be estimated from the observed DNL (EBM 13.1% and PN 14.9%) that as much as 53% to 63% of energy is transported in the plasma in the form of fatty acids in triglycerides. Thus, the requirement of calories from fat is high in growing LBW infants. Active lipogenesis is still necessary to meet the daily needs for energy, growth, and storage despite a high fat intake. These results indicate that LBW infants depend on DNL to supply the necessary energy, despite the fact that their intake of fat is within standard accepted clinical practice, and that milk formula containing <50% calories as fat may be suboptimal in this regard for the LBW infant.
The serum fatty acid profile for C16:0, C18:0, C18:1, and C18:2 in EBM-fed and PN subjects was not different despite existent differences in the fat content of EBM vs PN. This serum fatty acid profile is similar to previous reports in infants on breast-milk feedings.14 Such observations suggest that hepatic DNL in the newborn has an additional function in further modifying the dietary fatty acid profile to one that is physiologic within circulating plasma triglycerides. Because the process of DNL is thermogenically costly, we speculate that the conversion of the large amount of C18:2 to palmitate and stearate of triglycerides may worsen the energy deficit in the neonate receiving vegetable oil in intralipids from PN.17-19 Because it is more efficient to adjust fatty-acid composition by chain elongation and chain shortening rather than by DNL, the increased synthesis by DNL rather than by chain elongation of stearate in PN vs the EBM group is consistent with the lack of metabolic efficiency of dietary polyunsaturated fatty acids in the LBW neonates. We speculate that the infants in the EBM group may have gained more weight than the PN group because of efficient energy use.
Cholesterol content of EBM is much higher than PN, yet we observed cholesterol synthesis in growing LBW infants receiving breast-milk feedings similar to the PN group. The average fractional cholesterol synthesis (12.7% ± 2.1% in EBM and 17.4% ± 4.6% in PN) we observed is much higher than previously reported. In full-term infants receiving breast-milk feedings at <4 months of age, cholesterol synthesis was assessed to be 2.62% ± 0.38%.20 Hence, cholesterol synthesis in growing LBW infants may be dependent on tissue synthesis and modulated by gestational age, in addition to the dietary cholesterol intake.
In summary, hepatic DNL in LBW infants was assessed using [2-^sup 13^C] sodium acetate enrichment and mass isotopomer distribution analysis. In contrast with healthy adults, de nova synthesis of palmitate, stearate, and cholesterol from acetyl CoA is found to be very active in stable LBW infants receiving standard nutritional energy intake of carbohydrate and fat by either breast-milk feedings or IV fat emulsion in PN. Thus, LBW infants maintaining adequate postnatal growth may still need additional fat energy when provided with the standard acceptable intake of fat.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the nursing, pharmacy, and administrative staff at UCLA NICU and at Santa Monica NICU for their enthusiastic participation in research. The study was partly supported by UCLA, GCRC grant number MO1-RR00865. The Stable Isotope Laboratory at Harbor UCLA is supported by National Institutes of Health grant MO1 RR0425 to the Harbor-UCLA General Clinical Research Center and PO1-CA 42710 to the UCLA Clinical Nutrition Research Unit.
