Main

Due to gastrointestinal immaturity, VLBW neonates (birth weight <1250 g) are nourished primarily through the i.v. route(1), and provision of adequate nutrition can be difficult in the first few weeks of their life. VLBW infants therefore often undergo periods of negative energy balance. In addition, VLBW premature babies face a high risk of developing sepsis and necrotizing enterocolitis(2).

During severe stress, protein wasting occurs and is associated with increased morbidity and mortality from muscle wasting and impaired immune function(3). Preservation of body protein under such circumstances may have clinical importance. Although glutamine is the most abundant amino acid in the body and can be synthesized de novo in mammalian species, the muscle free glutamine pool can be depleted with stress, and replenishment of the glutamine stores is associated with improved nitrogen balance in such conditions(4–6). In addition, glutamine is extensively taken up in the adult splanchnic bed(7, 8) and may have a trophic effect on intestinal mucosa in animals(9) and stressed adult humans(10, 11).

Although glutamine may play an instrumental role both in nitrogen homeostasis and in maintenance of gut trophicity, the rates at which premature neonates produce endogenous glutamine and use exogenous glutamine are unknown, and the potential benefits of enteral glutamine supplementation have not been explored in this population.

The present study therefore used infusion of stable isotope-labeled glutamine and leucine to: 1) assess rates of endogenous glutamine production in VLBW neonates and determine whether these infants are able to synthesize glutamine; 2) quantitate how much of an enteral load of glutamine is extracted in their splanchnic bed; and 3) determine the effect of enteral glutamine on the metabolism of leucine, an essential amino acid.

METHODS

Materials. Purchased lots of L-[1-13C]- and L-[5,5,5-2H3]leucine (96.8% 13C, and 87.9%2 H3, respectively), and L-[1-13C]-, and L-[1,2,3,4,5-13C5]glutamine (99.9% 13C, and 94.9% U-13C, respectively; Cambridge Isotope Laboratories, Woburn, MA; and Tracer Technologies, Somerville, MA) were tested for chemical, isotopic, and optical purity by GCMS. Tracer solutions were prepared in sterile 0.45% saline and verified to be sterile (plate culture) and pyrogen free (Limulus lysate assay). Infusates were passed through a 0.22-μm filter and stored in sterile containers at 4°C for <24 h until used.

Patients. Written consent was obtained from the parents of 11 VLBW neonates before enrollment, according to procedures approved by the Institutional Review Board of Shands Hospital, Gainesville, FL, the Nemours Children's Clinic Research Committee, and the Institutional Review Committee at Baptist Medical Center, Jacksonville, FL. Patients were recruited from Shands Hospital's neonatal intensive care unit. Patients were excluded if they had any congenital anomaly of the gastrointestinal tract, necrotizing enterocolitis, major surgery, were considered potentially nonviable, or if the mother elected to breast-feed before entering the study. The patients recruited for the current isotope infusion study were a subset of a larger group recruited for a study involving a total of 71 infants and designed to assess the effect of a 30-d enteral glutamine supplementation on the clinical outcome of VLBW infants(12).

Experimental design. Eleven premature infants whose mothers decided to formula feed were randomly assigned to the glutamine-supplemented Similac Special Care (Ross Products Division, Abbott Laboratories, Columbus, OH) or nonsupplemented Similac Special Care group. In both groups, PN with glucose, amino acids, and lipids began on d 3 of life and was advanced incrementally to approximately 85 nonprotein kcal·kg-1·d-1. The PN amino acids (Travasol, Baxter Health-Care Corporation) were started at 0.5 g· kg-1·d-1 and increased by 0.5 g·kg-1·d-1 increments until 3 g·kg-1·d-1 was achieved by d 8. Intralipid (Kabi Pharmacia, Clayton, NC) also began on d 3 and increased until 3 g·kg-1·d-1 was reached by d 13. Glucose was started at 5 mg·kg-1·min-1 and increased as tolerated to 12 mg·kg-1·min-1. The PN regimen did not include any glutamine.

Both groups received an enteral priming regimen similar to that reported previously(13), beginning on d 3. Diluted(half-strength Similac Special Care, 20 kcal/oz) formula, with or without supplemented glutamine, was fed during the priming phase (3-14 d of age). Powdered glutamine (Ajinomoto, Teaneck, NJ) was mixed fresh daily by the hospital pharmacy. There were no discernable differences (e.g. odor or appearance) between the control and glutamine-supplemented formulas. The glutamine-supplemented group initially received 0.08 g·kg-1·d-1 of glutamine; this was increased to 0.2 g·kg-1·d-1 on the day of the study, and reached 0.3 g· kg-1·d-1 by d 13. Table 1 summarizes the babies' nutritional intake on d 10, i.e. on the day of the isotopic study. There was no obvious difference between the groups regarding the number of patients with proven or suspected sepsis, nor their ventilatory status (Table 1).

Table 1 Patients' characteristics and parenteral intake on the day of the study
Full size table

Protocol for isotope infusion. The isotopic study was performed on d 10 of life, in the fed state, while babies were receiving both continuous i.v. nutrition and continuous enteral feeding with either supplemented(n = 6) or nonsupplemented (n = 5) Similac Special Care. All patients therefore had an OG tube and a central venous catheter in place for several days before the isotope infusion. Because some of the babies studied had been on assisted ventilation, they had an arterial line in place as well. In the babies who did not have an arterial line in place at the time of the study, a butterfly needle was inserted in a hand vein for sampling of arterialized-venous blood(14).

At 0730 h on the isotopic study day, a baseline arterial blood sample (0.5 mL) was obtained for measurement of background isotopic enrichment in plasma KIC and glutamine. Continuous infusion of the labeled leucine and glutamine was started at 0800 h and continued for 4 h until 1200 h through both the i.v. and the OG routes: L-[1-13C]Leucine and L-[1-13C]glutamine were infused through the OG tube (at rates of 16.3 ± 3.0 and 69.1 ± 13.5 μmol·kg-1·h-1, respectively), whereas L-[5,5,5-2H3]leucine and L-[13C5]glutamine were infused through the central venous catheter (14.9 ± 2.8 and 27.0± 5.4 μmol·kg-1·h-1, respectively)(Fig. 1). Both the enteral and the i.v. tracer infusions were continued at a constant rate for 4 h.

Figure 1
figure 1

Isotopic model used to assess glutamine metabolism in infants receiving OG glutamine. The i.v. infused [U-13C]glutamine served to quantitate overall glutamine Ra; the apperance of the enterally infused [1-13C]glutamine into plasma allowed us to quantitate the fraction of OG glutamine (1-f) that escapes first-pass extraction in the splanchnic bed (gut and liver). The fraction of glutamine Ra not accounted for by bioavailable, exogenous glutamine nor release from proteolysis (B), is attributed to glutamine de novo synthesis (D).

Full size image

Half-milliliter blood samples were drawn from the arterial catheter for determination of plasma concentrations and enrichments of KIC and glutamine after 200, 220, and 240 min of tracer infusion. The total volume of blood sampled was therefore ≅ 2 mL, which is <5% of blood volume in a 1,000-g baby.

Analytical methods. Known amounts of L-[3,3,3,4,4,5,5-2H7]leucine ([2H7]leucine), D,L-[3,4,5-2H3]glutamine, ([2H3glutamine]), and[3,3,3,4,4,5,5-2H7]α-ketoisocaproate([2H7]KIC) (Cambridge Isotope Laboratories, Woburn, MA) were added to each 100-μL aliquot of plasma to serve as internal standards for measurement of glutamine, leucine, and KIC concentration by reverse isotope dilution. Plasma glutamine was isolated and derivatized as described previously(15). KIC was isolated from 100 μL of plasma by passing the acidified plasma sample over an AG50 cation exchange column. To each KIC-containing fraction, 3 drops of 10 M NaOH and 200 μL of 0.36 M hydroxylamine HCl were then added, and samples were incubated at 60°C for 30 min to produce an oxime derivative. Samples were then cooled immediately on ice, acidified with 2 M HCl, and spiked with 1 mL of supersaturated ammonium sulfate. KIC was extracted twice by shaking after adding 8 mL of ethyl acetate. The supernatant was then dried under nitrogen. Each dry sample was spiked with 50 μL of N-methyl-N-(t-butyldimethylsilyl)-trifluoroacetamide and incubated for 24-36 h at room temperature to obtain an oxime-t-butyldimethylsilyl KIC derivative. This modified method enhanced the sensitivity of the KIC assay, allowing for the use of smaller volumes of plasma than the previously described t-butyldimethylsilyl derivative(16).

Isotopic enrichments in plasma leucine, KIC, and glutamine were determined by selected ion monitoring GCMS (Hewlett-Packard MSD 5970; Palo Alto, CA). For KIC, ions at m/z = 316, 317, 319, and 323, representing the prominent ions of natural KIC, [13C]KIC, [2H3]KIC, and[2H7]KIC, respectively, were selectively monitored. For glutamine, ions at m/z = 172, 173, 175, and 177, for natural,[1-13C]-, [2H3]-, and [13C5]glutamine, respectively, were monitored. The concentration of leucine and glutamine in plasma (μM) was measured using GCMS by reverse isotope dilution of the[2H7]leucine and [2H3]glutamine internal standards, respectively.

Isotope model and calculations. The dual isotope model used to assess glutamine kinetics under fed conditions is depicted in Figure 1.

Intravenous infusion of [13C5]glutamine enabled us to quantitate total glutamine appearance (RaGLN,μmol·k-1·h-1) into the plasma compartment(17): Equation (1) where Tiv is the rate of [13C5]glutamine tracer i.v. infusion (μmol·kg-1·h-1), and Eiv and Epv are the[13C5]glutamine enrichments (mol% excess) in the i.v. infusate and in plasma, respectively.

The simultaneous infusion of [1-13C]glutamine via the OG route was designed to trace the appearance of enterally infused glutamine into the systemic circulation. Because [1-13C]glutamine entering systemic plasma is diluted by the whole body glutamine turnover rate, the appearance of dietary [1-13C]glutamine into systemic plasma was estimated as:Equation (2) where Epg is the[1-13C]glutamine enrichment in plasma at steady state (mol% excess).

The ratio of dietary [1-13C]GLN Ra to the actual amount of [13C]glutamine infused into the OG tube reflects the fraction of enteral [1-13C]glutamine that escapes splanchnic uptake:Equation (3) where Tog and Eog are the rate of OG [1-13C]glutamine infusion(μmol·kg-1·h-1) and the 13C-enrichment(mol% excess) of the [1-13C]glutamine tracer infused through the OG tube; and FSUGLN is the fraction of enterally administered glutamine that undergoes first-pass uptake in the splanchnic bed.

The amount of unlabeled glutamine appearing into plasma from the enteral perfusion will therefore be equal to (1 - FSUGLN) × (rate of unlabeled enteral glutamine infusion). The rate of infusion of natural, unlabeled glutamine was calculated as the sum of the free, unlabeled glutamine added to the formula as a powder, plus the estimated contribution of the bound glutamine residues present in the formula (assuming the protein in the formula contains 6.95 g of bound glutamine per 100 g of protein; see below).

Because leucine is an essential amino acid, the endogenous leucine flux (e.g. total leucine Ra minus the contribution from exogenous nutrition) is entirely derived from protein breakdown (BLEU). In contrast, the nonessential amino acid glutamine has, in addition to exogenous supply, two inflow components to its endogenous flux: 1) release from protein breakdown(BGLN), and 2) glutamine de novo synthesis (DGLN). The former is estimated as follows:BGLN = k × BLEU, where k is the assumed ratio of glutamine to leucine content of body protein.

There was no parenteral source of natural glutamine in the patients. Thus, the overall glutamine appearance rate(RaGLN) is described by the following equation:RaGLN = [(1 - FSUGLN) × OGGLN]+ BGLN + DGLN (Fig. 1)(8).

Whereas the leucine content and glutamine + glutamate content are known to be 8.0 and 13.9 g of amino acid/100 g of protein, the relative contribution of glutamine per se to the total remains to be defined. For estimating BGLN in previous studies, we assumed that glutamine residues contributed 13.9 g/100 g of protein(17, 18). However, because the glutamate content of whole body protein must be different from zero, this results in underestimates of DGLN. Therefore, in accordance with more recent work(7, 8), we assumed in the present study that glutamine residues contributed one-half of the total glutamine + glutamate i.e. 6.95 g/100 g of protein.

Equations analogous to Equations 1-3 were used for calculating leucine Ra(RaLEU), the amount of leucine arising from enteral leucine delivery, and leucine splanchnic uptake, based on plasma[2H3]KIC and [13C]KIC enrichments at steady state. Because leucine kinetics were measured under fed conditions, both exogenous and endogenous leucine contributed to leucine Ra (flux). In addition, two sources contributed exogenous leucine: indeed, PN and the fraction of dietary (orogastrically infused) leucine that escaped splanchnic uptake. The rate of leucine release from protein breakdown (i.e. endogenous leucine flux) was therefore calculated by subtracting both PN leucine intake and enteral leucine delivery-corrected for the fraction of enterally supplied leucine that underwent splanchnic uptake-from total leucine Ra.

Statistics. Results are expressed as means ± SD. Comparisons between baseline conditions and infusion of enteral glutamine were performed a using nonparametric test (Mann-Whitney).

RESULTS

As shown in Table 1 gestational age, birth weight, and parenteral intake of energy and amino acids did not differ between the glutamine-supplemented group and the control group.

There was no difference between the groups in terms of overall leucine Ra, PN leucine supply, or OG leucine intake (Table 2; Fig. 2). The fraction of leucine undergoing first-pass splanchnic extraction was slightly, but not significantly, higher in the glutamine-supplemented group, compared with controls. The estimated rate of leucine release from protein breakdown was slightly, but not significantly, lower in the glutamine-supplemented group, compared with controls.

Table 2 Leucine kinetics(μmol·kg-1·h-1) in VLBW neonates receiving PN and OG infusion of a glutamine-supplemented versus a nonsupplemented formula
Full size table
Figure 2
figure 2

Leucine Ra (triangles) and leucine release from protein breakdown (B; circles) in VLBW infants receiving PN, along with OG infusion of a glutamine-supplemented formula(closed symbols) or a nonsupplemented formula (open symbols).

Full size image

As shown in Table 3, and as per protocol design, dietary glutamine intake was higher in the glutamine-supplemented babies than in controls. Plasma glutamine concentration, did not, however, achieve higher values in the supplemented groups. There was a trend toward lower rates of glutamine de novo synthesis, in the glutamine-supplemented group: none of the kinetic parameters measured differed between the groups (Fig. 3).

Table 3 Glutamine kinetics(μmol·kg-1·h-1) in VLBW neonates receiving PN and OG infusion of a glutamine-supplemented versus a nonsupplemented formula
Full size table
Figure 3
figure 3

Glutamine Ra (triangles) and contribution of proteolysis (B; circles) and glutamine de novo synthesis (D; half-circles) in VLBW infants receiving PN, along with OG infusion of a glutamine-supplemented formula (closed symbols), or a nonsupplemented formula (open symbols).

Full size image

DISCUSSION

To our knowledge, this study is first to quantitate the rates of whole body glutamine production and utilization in VLBW neonates. We observed that, by the 10th d of life, premature neonates are capable of substantial glutamine de novo synthesis, and that glutamine is extensively extracted from the enteral lumen. The low doses of glutamine supplied enterally failed to affect whole body leucine flux.

Even though glutamine failed to affect leucine Ra, an index of protein breakdown, the present data do not rule out a potential anabolic effect of glutamine in this population, for several reasons.1) Rates of protein synthesis were not measured in the current study, and the anabolic effect of glutamine is more likely due to an increase in protein synthesis rather than a decrease in protein breakdown. Indeed, in healthy adults, enteral glutamine infusion enhanced nonoxidative leucine disposal, an index of whole body protein synthesis, with no change in leucine Ra(19). In the current study, the simultaneous infusion of several 13C-labeled species prevented accurate determination of leucine oxidation and, consequently, nonoxidative leucine disposal. 2) The small number of patients and wide dispersion of results (Fig. 2) may have precluded the detection of a reduction in proteolysis. 3) The dose of glutamine-0.2 g·kg-1·d-1-was lower than those (e.g. 0.5 g·kg-1·d-1) shown to enhance nitrogen(5, 6) or leucine(19) balance in adults. 4) Finally, a local anabolic effect on the gut cannot be excluded. In VLBW infants, glutamine supplementation decreased the incidence of sepsis from enterally derived bacteria(12) and improved tolerance to oral feeding(12, 20): both effects may reflect a trophic effect of glutamine on gut mucosa. In the current study, however, glutamine failed to increase leucine extraction in the splanchnic bed, thus providing no evidence for increased leucine utilization for protein synthesis in splanchnic tissues.

The glutamine turnover rates observed here are comparable to those observed in healthy 13 ± 4-mo-old infants(17). De novo synthesis accounted for ≅83% and ≅84% of endogenous glutamine appearance in the glutamine-supplemented and control VLBW infants, respectively. Admittedly-as discussed above- rates of glutamine de novo synthesis are only estimates; regardless of the assumptions made in the calculations, the contribution of de novo synthesis nevertheless largely exceeds that of proteolysis. Indeed, even if we assumed glutamine content of body protein to be 13.9 g/100 of g protein, a conservative estimate of glutamine de novo synthesis would account for ≅67% of glutamine Ra. The relative contribution of de novo synthesis to overall glutamine flux is thus of the same magnitude as in adults(8, 18) and older children(21). This demonstrates that by the 10th d of life, and despite stress, glutamine de novo synthesis is fully active in VLBW premature neonates. Because neither rates of glutamine oxidation nor the size of muscle free glutamine pool were assessed in the present study, it is unclear whether these rates of glutamine de novo synthesis meet the requirements of glutamine-consuming tissues. Although comparisons between different populations should be made with caution, it is of interest to note that-consistent with studies by others(20)-the VLBW infants in the current study had lower plasma glutamine concentrations than healthy adults(8, 18), prepubertal children(21), and older infants(17), studied with the same methodology, e.g. 330 ± 116 μM in VLBW in current study versus 536 ± 119 μM (p< 0.05), in 13 ± 4-mo-old infants(16). Thus glutamine metabolic clearance rate (Ra/glutamine concentration) was notably higher in the VLBW babies: ≅1.9versus 1.3 L·kg-1·h-1. Although the significance of this metabolic clearance rate is unclear, we speculate that the glutamine requirement of splanchnic tissues may be higher in VLBW babies, leading to lower circulating glutamine levels.

The measurement of the rate of first-pass splanchnic glutamine extraction assumes that: 1) molecules of the enterally infused glutamine tracer that fail to reach systemic blood are used in the splanchnic tissues;2) free glutamine is quantitatively released from hydrolysis of formula protein in the intestinal lumen; and 3) glutamine is quantitatively absorbed through the luminal membrane of the intestinal epithelium. Although the absorption of glutamine has not been evaluated in newborn babies, extensive absorption of leucine has been documented in that population(22), and it is fair to assume that glutamine absorption is at least as fast, because glutamine absorption is highly effective in adults(23). The rate of splanchnic glutamine extraction did not differ from that documented in adults(7, 8), suggesting that the gut is a prominent“target organ” of glutamine even in the first few days of life. The fractional rate of extraction was the same in both groups, despite the≅10-fold difference (≅49 versus 5μmol·kg-1·h-1) in the load of glutamine delivered into the gut; this suggests that glutamine extraction had not reached saturation in the glutamine-supplemented group.

In summary, even though plasma glutamine levels tend to be lower in premature neonates than in older, term infants(17), VLBW neonates are capable of producing endogenous glutamine through de novo synthesis. By the 10th d of life, the fractional rate of glutamine extraction in their splanchnic bed is similar to that of adults. The low doses of glutamine used did not inhibit protein breakdown, as measured at a whole body level. Further studies would have to be designed to determine whether glutamine alters rates of protein synthesis in premature infants.