Main

Damage to the surfactant system has been implicated in the pathophysiology of the acute lung injury which occurs in humans in a wide variety of acute processes such as lung infection, sepsis, trauma, ARDS, or neonatal RDS, or occurs secondary to mechanical ventilation. All of these processes lead to a widespread proteinaceous edema within the lungs and to a release of inflammatory mediators that could maintain lung inflammation (1–6).

Exogenous surfactant has been advocated as therapy in humans to overcome respiratory failure by restoring more uniform mechanical properties of the lungs. This approach has been proven to be beneficial in neonates with RDS (7), but the results in adults have been less encouraging (5,8,9). No data are available to date in humans on the endogenous surfactant synthesis and turnover in normal and injured lungs to support the use of exogenous surfactant after the neonatal period.

We have recently published that small preterm infants with RDS have a low fractional synthesis of lung PC palmitate from plasma glucose (10). Whether this low incorporation is caused by the precursor molecule used or the severity of the lung immaturity of the infants studied remains to be investigated. In the present study, we tested whether the labeling of plasma FFA with stable isotopes is a suitable technique for the measurement of surfactant synthesis in young infants. We developed a novel method for the measurement of endogenous surfactant production and turnover using a constant i.v. infusion of the albumin-bound [U-13C]PA and [U-13C]LLA. With this method, we saw a significant incorporation of the tracers into surfactant PC, and new data on endogenous surfactant synthesis and turnover in humans could be obtained.

METHODS

Surfactant synthesis was studied in eight infants whose clinical characteristics are reported in Table 1. All patients were admitted to the Neonatal or Paediatric Intensive Care Units of the Department of Paediatrics, University of Padua, Italy for respiratory failure. The inclusion criteria were 1) age less than 6 mo, 2) stable hemodynamic condition during the isotope infusion, and 3) respiratory failure that required endotracheal intubation for at least 48 h. Infants were excluded from the study when they presented with signs of liver and renal failure, seizures, or if they received blood products during the study. All infants had arterial and central venous lines placed for clinical monitoring and no one received intravenous lipid emulsion during the study. Feeding was started gradually, 3 d after the start of infusion. All the medications given to the patients during the study were recorded. The study protocol was approved by the Ethical Committee on Human studies at the Department of Paediatrics of the University of Padua and was conducted according to the principles expressed in the Declaration of Helsinki. Written informed consent was obtained from the parents.

Table 1 Clinical characteristics of the infants studied
Full size table

Study protocol. [U-13C]PA (16:0; purity: 99%) and [U-13C]LLA (18:2n-6; purity: 97%) were purchased from Martek Biosciences (Columbia, MD). Chemical and isotopic purity was confirmed by gas chromatography mass spectrometry. [U-13C]PA and [U-13C]LLA were bound to human albumin (Merieux, Pasteur, Lyon, France) and prepared for i.v. infusion as described previously (11). The albumin-bound tracers were infused i.v. at a constant rate by a high precision calibrated syringe pump (M22, Harvard Apparatus Co, Inc. Natick, MA) over a 24-h period. The rates of infusion were 1.0 ± 0.26 and 1.19 ± 0.23 µmol·kg-1·h-1 for PA and LLA, respectively.

We used the arterial line for blood sampling and the venous line for tracer infusion. Before and during the isotope infusion, 0.6 mL of blood was drawn at 0, 3, 6, 12, 18, and 24 h to determine the isotopic enrichment of PA and LLA enrichment in plasma FFA. The blood drawn was placed in tubes containing EDTA and immediately centrifuged at 1 300 × g; plasma after separation was stored in tubes containing pyrogallol as antioxidant at -20°C until analysis.

Tracheal aspirates were obtained before the start of the isotope infusion and every 6 h thereafter until the patient was extubated. Twenty seconds after 0.5-mL normal saline was injected into the tracheal tube, suction was done beyond the tip of the endotracheal tube. Normal saline was added to the tracheal aspirate to a total of 3 mL. This was centrifuged at 150 × g for 10 min, and supernatant was stored at -20°C.

Analytical methods. Lipids were extracted from tracheal aspirates according to Bligh and Dyer (12). Surfactant PC was separated from other surfactant phospholipids by thin layer chromatography (13). The PC was derivatized by adding 2 mL of 3 M dry HCl methanol (14), and the fatty acids were extracted with hexane and stored at -20°C until analyzed. No tracheal aspirate contained visible blood.

Plasma lipids were processed as previously described (11). Plasma was delipidated according to Folch (15), lipid classes separated by thin layer chromatography, and their fatty acids derivatized as methyl esters. The quantification of plasma FFA and of FA of surfactant PC were performed as previously described (11).

The enrichment of individual fatty acids of surfactant PC was measured by gas chromatography isotope ratio mass spectrometry (VG Isotech, Midlewich, Chesire, UK) (10). The enrichment was expressed as APE, which represents the increase in percentage of 13C atoms in total carbon dioxide from the combusted compounds above baseline enrichment (before isotope infusion). Enrichments were corrected for the contribution of unlabeled carbon atoms during derivatization.

Isotopic enrichments of PA and LLA of plasma FFA were carried out on a Fisons MD 800 gas chromatograph quadrupole mass spectrometer (Fisons, Milan, Italy), using the chemical ionization mode with isobutane (11). Each plasma sample was measured in duplicate.

Calculations. Surfactant in the type II cells and laying on alveolar surface were regarded as one pool, because animal studies showed that recycling is much faster than the de novo synthesis and clearance (16,17). The following surfactant PCs kinetic parameters were calculated using both PA and LLA:

Secretion time was defined as the time lag between the start of the [U-13C]PA and the [U-13C]LLA infusion and the appearance of the respective labeled fatty acid in the surfactant PC. The time of appearance of the enrichment in PC-PA and PC-LLA was calculated by plotting the regression line for the linear increasing part of the enrichment versus time curve and extrapolating it to baseline enrichment (10,18).

FSR of PC-PA and PC-LLA is expressed as the percentage of the total surfactant PC pool synthesized from the respective plasma fatty acid per h. It is calculated by dividing the slope of the linear increase of the enrichment of PC by the plasma steady state enrichment of the respective FA of the plasma FFA pool (10,19).

Half-life of PC was calculated by exponential curve fitting at the final monoexponential part of the down-slope of the enrichment versus time curve.

Peak time is the time of maximum enrichment of surfactant PC after the start of the isotope infusion.

Data were expressed as individual values, group mean ± SD, and range.

RESULTS

In all infants, [U-13C]PA and [U-13C]LLA enrichments of plasma FFA reached steady state within 3 h from the start of the i.v. isotope infusion, and the slope of the enrichment curve over time did not deviate significantly between t = 3 h and t = 24 h (Fig. 1). Total FFA, PA, and LLA plasma concentrations and mol% values of 7 of the 8 infants studied were reported elsewhere (11).

Figure 1
figure 1

Isotopic enrichment (APE) of plasma free PA (A) and LLA (B) in eight critically ill infants during a 24 h constant i.v. infusion of albumin-bound [U-13C]PA and [U-13C]LLA. All infants (individual patients are identified by different symbols) reached steady state after the third h of the isotope infusion.

Full size image

The mean ± SD percentage values of PA and LLA in tracheal aspirate PC were 64.4 ± 10.6 and 2.0 ± 1.1 mol%, respectively. PC-PA and PC-LLA molar percentages were rather stable in the individual patients with a variability of less than 5% during the study period.

Figures 2 and 3 show the time curves of the 13C enrichment of individual fatty acids of surfactant PC in patient 7. Other fatty acids, besides PA and LLA, became enriched in surfactant PC. Figure 2 shows the nonessential fatty acids derived from the desaturation and/or chain elongation of labeled PA, namely 16:1n-7, 18:0 and 18:1n-9. Figure 3 shows 13C enrichment of LLA and of its metabolic derivatives, namely 20:3n-6 and arachidonic acid (20:4n-6).

Figure 2
figure 2

Isotopic enrichments above baseline (APE) of PA and its metabolic derivatives incorporated in surfactant PC after a 24 h i.v. infusion of [U-13C]PA in a patient with respiratory failure.

Full size image
Figure 3
figure 3

Isotopic enrichments above baseline (APE) of LLA and its metabolic derivatives incorporated in surfactant PC after a 24 h infusion of [U-13C]LLA in a patient with respiratory failure.

Full size image

Kinetic data calculated for each individual patients are shown in Table 2; FSR, secretion time, peak time, and half-life were calculated using both PA and LLA. Surfactant PC-PA and PC-LLA half-lives were calculated in five patients (patients 3, 5-8) because patients 1 and 4 were extubated within 72 h of the beginning of the study and patient 2 died 80 h after the beginning of the study. FSR for PA and LLA ranged from 0.8 to 3.4 and 0.5 to 3.8% per h, respectively, which correspond to mean values of 34 and 50% per d, respectively. We could not find any correlation between PA and LLA FSR and their respective PA and LLA mol% values in tracheal aspirate PC (data not shown). Mean secretion time and time to peak for PA versus LLA were similar (8.7 ± 4.9 versus 10.0 ± 7.2 h and 49.2 ± 8.9 versus 45.6 ± 19.3 h, respectively). Half-lives of PC-PA and of PC-LLA showed large differences in the individual patients (range: 16.8-177.7 h). In patient 5, half-lives of PC-PA and of PC-LLA were 177.7 versus 43.6 h, respectively. In patient 6, we found the reverse, and values were 33.2 versus 144.4 h, respectively.

Table 2 Surfactant phosphatidylcholine kinetics after a 24-h intravenous infusion of [U-13C]PA and [U-13C]LLA in critically ill infants
Full size table

DISCUSSION

In this study we measured the endogenous surfactant production and turnover in vivo in human infants by using stable isotope labeled fatty acids. We infused [U-13C]PA and [U-13C]LLA in infants with respiratory failure and measured the 13C-enrichment of selected fatty acids in surfactant PC.

Surfactant synthesis has been studied in vivo in animals by using several metabolic precursors of the surfactant PC molecule such as glycerol (20), glucose (21), choline (22), and more frequently, palmitic acid (17,23). After i.v. administration, these precursor molecules have different volumes of distribution and are subject to a different oxidative and nonoxidative metabolism before being taken up by the type II cells in the lungs.

We have recently infused [U-13C]glucose in preterm infants with RDS and measured the isotopic enrichment of PA in surfactant PC (10). By infusing labeled glucose as a tracer, we measured the PC-PA synthesized de novo from glucose via lipogenesis and incorporated into surfactant PC by the type II cells. After i.v. [U-13C]glucose, the site of PA synthesis can be the type II cell with direct incorporation of PA into PC-PA or the liver with the subsequent production of very low density lipoproteins containing labeled PA. Plasma lipids are taken up by the type II cells for PC synthesis (24). Under these experimental conditions, total surfactant PC synthesis could not be calculated, because PA from sources other than lipogenesis from glucose contribute to the synthesis of PC-PA.

In the present study, by infusing labeled PA, we measured the surfactant PC synthesized by the type II cells using plasma PA as metabolic precursor, but we could not calculate the surfactant PC-PA from de novo synthesis from glucose in the type II cells. Our patients also received an infusion of LLA, which is an essential fatty acid and cannot be endogenously synthesized. We used LLA to estimate the total PC surfactant production.

The incorporation of labeled PA and LLA in surfactant PC, namely secretion time, began 9 h after the start of the infusion. This does not suggest a different processing of PA versus LLA during the de novo synthesis of surfactant PC, which is consistent with in vitro studies (25). The secretion time corresponds to the time required by the alveolar type II cells to take up the fatty acids from plasma lipids and to process and excrete them as PC into the alveolar space in addition to the time required by surfactant PC to reach the tracheal bronchial tree. Slightly shorter times have been reported in 10-d-old lambs that received 3H-palmitate as a bolus (17).

We reported longer secretion times (19.4 ± 2.3 h) in preterm infants with RDS treated with exogenous surfactant who received an i.v. infusion of 13C-glucose as precursor for surfactant PC palmitate (10). Whether this difference is determined by the type of precursor used (fatty acids versus glucose), the different patient characteristics, (critically ill infants versus small preterm infants with RDS), or the dilution of the labeled PC that could have occurred via recycling in the type II cells after exogenous surfactant in the preterm infants with RDS is currently being studied. The time of maximum isotopic enrichment of the fatty acids in surfactant PC was approximately 2 d for both PA and LLA; these figures are comparable with the animal data in newborn lambs and rabbits (17,23,26).

Although the patients studied constitute a very heterogeneous group, the incorporation rate of PA in surfactant PC was consistently lower than the respective incorporation rate of LLA in surfactant PC. The mean incorporation rate of PA in surfactant PC was 34.2 ± 24.8% per d of the PC PA pool size, which is approximately one half of the incorporation of LLA (50.8 ± 26.0% per d). Therefore, the infants in the study synthesized one third of their surfactant PC PA from plasma PA and one half of their surfactant PC LLA from plasma LLA each d. These values did not correlate with the surfactant PA and LLA mol% values. Lower FSR for PA than for LLA could be explained by the de novo synthesis of palmitate from other substrates than from fatty acids. In the case of LLA, this process cannot occur in animals because LLA is an essential fatty acid and there is no endogenous synthesis of LLA. In animals, surfactant PC palmitate, derived from plasma fatty acids, represents only approximately 50% of the total PA incorporated in surfactant PC, whereas the other 50% is derived from glucose (21,27), lactate, ketone bodies, and glutamate (21,28). Despite that LLA represents only 2-3% of the total fatty acids of surfactant PC, its incorporation rate could be a more reliable measure of the rate of surfactant PC synthesis than that of PA. Therefore, according to this estimate, approximately 50% of the surfactant PC pool was synthesized on average by our group of patients in 1 d, which is in agreement with data in rabbits (17,27).

The half-life of surfactant PC calculated from the down slope of the PA time enrichment curve ranged between 1 and 7.4 d, with infants with the most severe respiratory failure having the shortest PA half-lives (patients 3, 6, and 7). The limited number of infants studied doesn't allow us to draw any firm conclusion, but the wide range suggests a different behavior of surfactant kinetics in our patients. In term newborn rabbits that received a bolus of [3H] PA i.v., the half-life of surfactant PC yielded comparable results ranging between 2 and 5.7 d (17). Furthermore, the half-lives of PC-PA and of PC-LLA exhibited rather different values in the individual patients. In patient 5, the half-lives of PA and LLA were 177.7 versus 43.6 h, respectively; in patient 6, we found the reverse and values were 33.2 versus 144.4 h, respectively. We do not have a clear explanation for these finding because most of the animal work has focused on the metabolism of disaturated PC and little attention has been paid to the metabolism of polyunsaturated fatty acids. Long half-lives of PC-PA could result from a reduced surfactant PC catabolism and/or could be associated to an "active" recycling. The latter process is known to exhibit a strong preference for saturated fatty acids (29). Also, longer half-lives of PC-LLA in PC could represent an inhibition of remodeling of unsaturated PC to saturated PC (25). A larger proportion of surfactant PC with palmitate, synthesized via de novo lipogenesis from glucose, could also contribute to shorter half-lives during periods of increased surfactant synthesis. We have recently observed increased incorporation of palmitate from glucose in preterm baboons who have been treated prenatally with corticosteroids in comparison with baboons who did not receive steroids (unpublished data).

In summary, our study describes a new and safe method for the measurement of surfactant synthesis and turnover in vivo in humans. These preliminary results in critically ill infants are consistent with previous data obtained from animal experiments and suggest that surfactant synthesis is a slow process. The use of stable isotope labeled fatty acids such as LLA and PA and of labeled glucose will expand our understanding of surfactant metabolism in humans. This will improve the rational use of exogenous surfactant in infants with lung injury (8,9).