点击显示 收起
1 From the Tropical Metabolism Research Unit (AB, MR, and TF) and the Section of Radiology (DS), University Hospital of the West Indies, University of the West Indies, Mona, Kingston, Jamaica, and the US Department of Agriculture/Agricultural Research Service, Department of Pediatrics, Children's Nutrition Research Center, Baylor College of Medicine, Houston, TX (FJ)
2 Supported by the US Department of Agriculture's Agricultural Research Service under Cooperative Agreement no. 58-6250-6001, NIH grant no. 1RO1 DK56689, and a grant from The Wellcome Trust. 3 Reprints not available. Address correspondence to F Jahoor, Department of Pediatrics, Children's Nutrition Research Center, Baylor College of Medicine, 1100 Bates Street, Houston, TX 77030-2600. E-mail: fjahoor{at}bcm.tmc.edu.
ABSTRACT
Background: Fatty infiltration of the liver is associated with an increased morbidity and mortality in children with severe protein-energy malnutrition (PEM), but its pathogenesis remains unclear. Although impaired synthesis of VLDL apolipoprotein B-100 (VLDL-apo B-100) is generally accepted as the pathogenetic mechanism, the rate of it synthesis has not been measured in children with PEM.
Objective: The objective of the study was to ascertain the relation between the degree of hepatic steatosis and the rate of VLDL-apo B-100 synthesis in children with PEM.
Design: The fractional and absolute rates of VLDL-apo B-100 synthesis were measured with a prime-constant intravenous infusion of [2H3]leucine in 13 severely malnourished children (8 boys and 5 girls) aged 7–18 mo. Hepatic fat content was estimated by computerized tomography scanning by using the ratio of liver to spleen (L:S) attenuation. The ratio is inversely related to hepatic fat content such that the lower the L:S, the greater the amount of fat in the liver.
Results: There were significant inverse relations between L:S attenuation and VLDL-apo B-100 concentration (P < 0.02), the absolute rate of VLDL-apo B-100 synthesis (P < 0.02), and plasma triacylglycerol (P < 0.02) and serum cholesterol (P < 0.05) concentrations.
Conclusions: These results suggest that children with PEM synthesize VLDL-apo B-100 at a faster rate as the degree of hepatic fat infiltration increases. Thus, fatty infiltration of the liver in PEM is not due to a reduction in the synthesis of VLDL-apo B-100.
Key Words: VLDL apolipoprotein B-100 • VLDL-apo B-100 • protein-energy malnutrition • children • fatty liver
INTRODUCTION
Fatty liver is a common feature of children with protein-energy malnutrition (PEM), and a hepatic lipid content >40% of liver weight is associated with a very poor prognosis (1, 2). Although it is generally accepted that impaired synthesis and secretion of VLDL apolipoprotein B-100 (VLDL-apo B-100) is the primary underlying defect responsible for the excess triacylglycerol deposition in the livers of these children (2-6), there has never been any direct experimental evidence to support this understanding. The precise mechanism of the excess triacylglycerol deposition has therefore remained a topic of debate (7-9).
The main component of VLDL is triacylglycerol, but the lipoprotein also transports cholesterol, cholesterol esters, and phospholipids from the liver to other tissues of the body. The first proposal that impaired synthesis of the apolipoprotein moiety of VLDL was responsible for the excess deposition of triacylglycerol in the livers of malnourished children was based exclusively on clinical observations that children with PEM and fatty livers had lower serum ß-lipoprotein (ie, LDL) cholesterol and triacylglycerol concentrations at admission than during nutritional rehabilitation, when liver fat content was believed to start receding (4, 5, 9), and also in comparison with the values in well-nourished children (5). This, however, was not a consistent observation (6, 10-12). It was further suggested that the impaired synthesis of VLDL-apo B-100 was due to a shortage of amino acids because of the chronically inadequate dietary protein intake of children with severe malnutrition (2, 6, 9, 13). The argument for this proposal was based on observations of a slower rate of incorporation of radiolabeled glycine into the protein moiety of LDL and VLDL in rats fed a low-protein diet (13) and of the greater amount of radiolabeled oleic acid incorporated into plasma triacylglycerol after rats fed a protein-free diet were injected with a plasma protein fraction containing the apoprotein of LDL (13). However, the dietary protein provided to the rats in both studies was not comparable to the amount provided by any human diet, even including the poor diets that precipitate PEM. Although the evidence is not conclusive, to date the rate of VLDL-apo B-100 synthesis has not been measured directly in malnourished children. In the current study, we used a stable isotope tracer method to measure the rate of VLDL-apo B-100 synthesis in severely malnourished children who were receiving a 5% dextrose solution. We also investigated the relation between the rate of VLDL-apo B-100 synthesis and liver fat content as ascertained by computerized tomography (CT) scanning (14-16). Dietary protein was not given because we believed that it would not be easy to disentangle any possible independent effect of providing amino acids on the rate of VLDL-apo B-100 synthesis.
SUBJECTS AND METHODS
Subjects
Thirteen children were recruited from among those admitted to the Tropical Metabolism Research Unit (TMRU) of the University of the West Indies for treatment of severe malnutrition. There were 8 boys and 5 girls aged 7–18 mo who were diagnosed with kwashiorkor, marasmic kwashiorkor, and marasmus according to the Wellcome Classification (17). The physical characteristics of the subjects are shown in Table 1.
View this table:
TABLE 1. Diagnosis and anthropometric characteristics of the children at time of study1
Written informed consent was obtained from at least one parent of each child enrolled. The Faculty of Medical Sciences/University Hospital of the West Indies Ethics Committee and the Baylor Affiliates Review Board for Human Subject Research, Baylor College of Medicine, approved the study.
Resuscitative treatment
During hospitalization, the children were treated according to a standard treatment protocol that is based on an understanding of the metabolic state at different stages of treatment (18). Briefly, the initial phase was a resuscitative period of treatment from admission until appetite returned, edema was lost, and infection was cleared. During this period, fluid and electrolyte imbalances were corrected, infections were treated with broad-spectrum antibiotics, and the children were fed a milk (Nan; Nestlé SA, Vevey, Switzerland)-based resuscitative diet (protein: 5% of energy; carbohydrate: 74% of energy; fat: 21% of energy) that aimed to provide 417 kJ · kg–1 · d–1 and 1.2 g · kg–1 · d–1 of protein, which is adequate for maintenance of body weight. Feeds were offered as boluses every 3 h or as smaller boluses every 2 h if the child was having problems tolerating the larger volume of formula. The diet was supplemented with vitamins and a mineral mix to provide adequate amounts of micronutrients. This study was carried out during the acute resuscitation phase, but, on the day of the study, the subjects did not receive the usual milk feeds during measurements (see Study design).
Study design
The rate of VLDL-apo B-100 synthesis was measured with an intravenous infusion of isotopically labeled leucine a mean (±SEM) 4 ± 1 d after admission. The isotope infusions were performed over a 6-h period, starting 2 h after the last bolus meal. To avoid hypoglycemia during the infusion protocol, a 5% dextrose solution was infused intravenously at 3 mg · kg–1 · min–1 starting 2 h before the isotope infusion. A noncontrast abdominal CT scan was performed within 24 to 48 h of the isotope infusion.
Assessment of liver fat
Assessment of liver fat content was done by CT scan on the basis of previously established criteria that have been validated by liver biopsy and liver histomorphometry (14-16). Enrolled children underwent a noncontrast abdominal CT scan within 24 to 48 h of the infusion by using a Sytecsynergy scanner (General Electric Co, Fairfield, CT). The children were sedated with chloral hydrate (50 mg/kg) before each scan according to standard radiologic practice. To reduce radiation exposure, a single cross-sectional CT scan of 10-mm thickness was taken at the level of the intervertebral disc between the body of the 12th thoracic and 1st lumbar vertebrae to include both lobes of the liver, the renal cortex, and the spleen. For all scans, the window level and window width were kept constant, and the machine was operated in the tissue optimization mode. A region of interest (ROI) was placed on 4 areas of the liver and 1 area of the spleen: at depths of 1.5, 2.0, and 3.0 cm from the liver capsule on the right lobe of the liver, at a depth of 1.5 cm from the liver capsule on the left lobe, and at a depth of 1.5 cm from the splenic capsule on the spleen. Splenic attenuation was measured from the latter ROI. Care was taken not to include major portal, arterial, and venous vessels. For each ROI, the attenuation measured in Hounsfield units (HU) was recorded. The mean liver attenuation was calculated from the 4 liver ROIs, and the ratio of mean liver attenuation to spleen attenuation (L:S) was ascertained. L:S is inversely related to hepatic fat content such that the lower the L:S, the greater the amount of fat in the liver, and a ratio <1 denotes significant hepatic steatosis (16, 19, 20).
Tracer infusion protocol
The rate of VLDL-apo B-100 synthesis was measured over a 6-h period by intravenous infusion of a sterile solution of [2H3]leucine (Cambridge Isotope Laboratories, Woburn, MA). Two intravenous access sites were established in opposite arms by the insertion of 22- or 24-gauge catheters after preparation of the access sites with a topical anesthetic (EMLA cream; Astra Pharmaceuticals Ltd, Langley, United Kingdom). One intravenous catheter was used for infusion of the [2H3]leucine and the other for blood sampling. After a 2-mL venous blood sample was drawn for baseline measurements, a 10 µmol · kg–1 priming dose of [2H3]leucine was administered, and this was followed immediately by a continuous infusion of [2H3]leucine at 10 µmol · kg–1 · h–1 for 6 h. Additional 1-mL blood samples were drawn hourly during the infusion.
Sample collection and analysis
The blood samples were drawn into chilled tubes containing Na2EDTA and a cocktail of sodium azide, merthiolate, and soybean trypsin inhibitor. They were centrifuged immediately at 1000 x g for 15 min at 4 °C, and the plasma was removed and stored immediately at –70 °C for later analyses. Serum cholesterol was assayed with the use of the enzymes cholesterol esterase and cholesterol oxidase (Infinity cholesterol reagent; Sigma Diagnostics Inc, St Louis, MO). Plasma triacylglycerol concentrations were measured by the colorimetric method by using the combined reactions of lipase, glycerol kinase, and L--glycerol-phosphate oxidase (Johnson & Johnson Clinical Diagnostics Inc, Rochester, NY). VLDL was isolated from plasma by ultracentrifugation as previously described (21); its apo B-l00 was precipitated, isolated, and acid hydrolyzed; and the isotopic enrichment of bound leucine was measured by gas chromatography–mass spectrometry as previously described (22). Briefly, the amino acids released from the protein by acid hydrolysis were purified by cation exchange chromatography and converted to the n-propyl ester heptaflourobutyramide derivative, and leucine isotope ratio was measured by monitoring ions at a mass-to-charge ratio of 349 to 352 on a gas chromatography mass spectrometer (model 5988A; Hewlett-Packard, Palo Alto, CA). VLDL-apo B-l00 concentration was measured by radial immunodiffusion on an aliquot of the supernatant.
Calculations
The fractional synthesis rate (FSR) of VLDL-apo B-100 was calculated from the rate of incorporation of [2H3]leucine into the protein during the rise to a plateau and the isotopic enrichment of the protein at the plateau, as described by Lichtenstein et al (23).
RESULTS
Thirteen children, aged 7–18 mo, participated in the study. On the basis of the Wellcome Classification (17), 6 were diagnosed with marasmus, 3 with kwashiorkor, and 4 with marasmic-kwashiorkor. The physical characteristics of the subjects are presented in Table 1. All of the subjects were severely malnourished and had markedly lower than expected weight-for-age and weight-for-length. Biochemical measurements in the subjects are shown in Table 2. All subjects were anemic, and 11 were hypoalbuminemic. Plasma bilirubin concentrations were within the normal range in all of the subjects. Aspartate aminotransferase concentration, measured in 12 of the children, was above normal. Eleven of the 13 subjects had evidence of one or more infections at admission.
View this table:
TABLE 2. Biochemical characteristics of the children at hospital admission1
The actual mean intakes of energy and protein before the isotope infusion protocol were 89% of the goal, which reflects the anorexia that is characteristic of the early resuscitative phase of treatment (Table 3). There were no significant relations between mean protein or energy intakes before the experiment and VLDL-apo B-100 concentration, FSR, or ASR. Liver span, the liver and spleen CT attenuation numbers, and L:S attenuation are shown in Table 4. Five of the children had L:S <1, which indicated marked hepatic steatosis. There were significant inverse relations between L:S and the VLDL-apo B-100 concentration (P < 0.02; Figure 1), the VLDL-apo B-100 ASR (P < 0.02; Figure 2), plasma triacylglycerol concentrations (P < 0.02; Figure 3), and cholesterol concentrations (P < 0.05; Figure 4). There was no significant relation between L:S and the VLDL-apo B-100 FSR.
View this table:
TABLE 3. Dietary energy and protein intakes of children on the day before isotope infusion
View this table:
TABLE 4. Liver span and computerized tomography (CT) scans of liver and spleen of children1
FIGURE 1.. The association between VLDL-apolipoprotein B-100 (VLDL-apo B-100) concentrations and the ratio of liver attenuation to spleen attenuation in children with protein-energy malnutrition. CT, computerized tomography.
FIGURE 2.. The association between VLDL-apolipoprotein B-100 (VLDL-apo B-100) absolute synthesis rates (ASR) and the ratio of liver attenuation to spleen attenuation in children with protein-energy malnutrition. CT, computerized tomography.
FIGURE 3.. The association between plasma triacylglycerol concentrations and the ratio of liver attenuation to spleen attenuation in children with protein-energy malnutrition. CT, computerized tomography.
FIGURE 4.. The association between plasma cholesterol concentrations and the ratios of liver attenuation to spleen attenuation in children with protein-energy malnutrition. CT, computerized tomography.
DISCUSSION
The aim of this study was to ascertain whether the fatty liver seen in severely malnourished children is related to impaired synthesis of the apo B-100 moiety of VLDL. Our results showed that the children with more fat in their livers (ie, lower L:S) synthesized VLDL-apo B-100 at a faster rate than did children with less fat in their livers, and that this faster synthesis was associated with higher plasma concentrations of the lipoprotein. The children with more fat in their livers also had higher concentrations of plasma triacylglycerol and cholesterol than did the children with less fat in their livers, which suggested increased mobilization of lipids from the liver. These findings suggest that greater lipid deposition in the livers of children with severe malnutrition is not associated with impaired synthesis of VLDL-apo B-100, as was proposed previously (2-6).
In the current study, it was not considered ethical to perform a biopsy of liver tissue. CT scanning has emerged as an excellent noninvasive tool for detecting and quantifying hepatic fat deposition (16, 19, 20), because it shows a significant inverse relation between the hepatic fat content and the hepatic attenuation number. As fat accumulates in the liver, the tissue's absorption of the X-rays decreases, which leads to a decrease in the CT HU (14-16). The use of absolute CT attenuation to define hepatic steatosis is limited, however, because of variations of the number. This is due to the fact that the actual HU is affected by nonhepatic factors, such as body size and instrument variations (20, 26). Expressing liver attenuation in proportion to splenic attenuation (L:S), however, overcomes this limitation because the spleen acts as an internal control. The splenic attenuation number is normally 8–10 HU less than the liver attenuation number (27), and therefore L:S <1 is associated with significant hepatic steatosis (16, 19, 20).
The assembly of VLDL and its secretion from the liver is a two-step process. Initially, hepatic apo B-100 associates with lipids to form dense lipoprotein particles in the rough endoplasmic reticulum, and then it is further lipidated to VLDL, probably in the smooth endoplasmic reticulum (28). The general belief that the fatty liver of children with PEM resulted from impaired synthesis of the apoprotein moiety of VLDL was based largely on 2 lines of indirect evidence. The first was observations that the plasma triacylglycerol and cholesterol concentrations in children with severe malnutrition and fatty livers were low at admission and that they rose during treatment (3-6). Thus, the role of impaired VLDL-apo B-100 synthesis in the pathogenesis of the fatty liver of severely malnourished children was imputed from the assumption that an increase in plasma triacylglycerol and LDL-cholesterol concentrations indicated increased synthesis of all lipoproteins (2). However, this generally accepted hypothesis of impaired VLDL-apo B-100 synthesis was never based on any actual measurement of the rate of synthesis. Moreover, several other studies in children with severe malnutrition have found plasma triacylglycerol concentrations that are low (10, 11, 29), normal (10, 11, 29), or high (10-12, 29) at admission and that rise, remain unchanged, or fall, respectively, during treatment (10, 11, 29). Because of these widely varied values, higher plasma triacylglycerol concentrations during treatment cannot be used as a reliable indicator of impaired VLDL-apo B-100 synthesis before treatment.
The second premise linking the accumulation of fat in the liver with impaired synthesis of VLDL-apo B-100 in severely malnourished children was based on the argument that dietary protein deficiency resulted in a reduction in the rate of synthesis of the apolipoprotein. This argument, in turn, was based primarily on the findings of a slower incorporation of radiolabeled glycine into the apoprotein of LDL and VLDL in rats fed a low-protein diet and of a faster incorporation of radiolabeled oleic acid into plasma triacylglycerol when rats fed a protein-free diet were injected with LDL apolipoprotein (6, 13). However, the protein supplied to the rats in both of those studies was not comparable to the amount supplied by any human diet, including the poor diets that precipitate severe malnutrition. In the current study, the children were treated before the experiment with a resuscitative diet that aimed to provide 417 kJ · kg–1 · d–1 and 1.2 g · kg–1 · d–1 of protein. The actual intake was 11% lower, but the absence of any significant association between the amount of dietary protein and the rate of VLDL-apo B-100 synthesis does not support the hypothesis that dietary protein deficiency was the underlying cause of fat deposition in the livers of children with PEM. In addition, there was no significant association between VLDL-apo B-100 synthesis and the energy intake from the resuscitative diet before the studies, which suggested that it was unlikely that the overall resuscitative diet had an effect on apo B-100 synthesis.
Plasma triacylglycerol concentrations were significantly higher in the children with more liver fat than in those with less liver fat. Similarly, the amount of fat in the liver was positively related to the concentrations of VLDL-apo B-100 and serum cholesterol, which indicated enhanced removal of lipid from the livers of the children with greater degrees of hepatic steatosis. Our current findings seem to support the contention of others (7, 8, 12) that impaired VLDL synthesis is not the cause of excess triacylglycerol deposition in the livers of malnourished children. It may be argued that the significant regression between L:S and the VLDL-apo B-100 ASR and concentration (Figures 1 and 3) is dependent on 2 or 3 subjects; if they were excluded, then there would be no significant relation. In such a case, the data still would not support the widely held belief that hepatic steatosis in malnourished children is due to impaired synthesis of VLDL-apo B-100. One possible limitation of this study is that the plasma volume specific to each child was not used in the calculation of the absolute rate of synthesis. However, plasma volumes specific for kwashiorkor, marasmic-kwashiorkor, and marasmus were used, and the values were based on measurements that we have done in a group of malnourished children of similar age, at a similar stage of rehabilitation, and with similar fluid intake.
There is alternative convincing evidence to support the involvement of other mechanisms in the pathogenesis of the fatty liver of childhood malnutrition (7, 30-33). In theory, the amount of hepatic triacylglycerol available for export can exceed the rate of removal because of 1 or more of 4 possible mechanisms—either increased hepatic fatty acid (FA) synthesis; impaired FA oxidation by hepatocytes; increased hepatic FA influx secondary to a stimulated rate of lipolysis, impaired whole-body FA oxidation, or both; or impaired removal of triacylglycerol from the liver by VLDL. For example, Fletcher (30) proposed that hepatic FA availability was increased because of increased FA synthesis from glucose. This proposal was based on Fletcher's finding of a significantly lower glucose-6-phosphatase activity in the liver tissue of malnourished children with fatty livers than in the liver tissue of children who had recovered from malnutrition (30). In addition, 2 studies by Lewis et al (31, 32) reported faster plasma palmitate flux and higher FA concentrations, which were indicative of a stimulated rate of lipolysis, in malnourished children with fatty livers than in well-nourished children. An increased lipolysis together with the finding that whole-body lipid oxidation is markedly slower in children with kwashiorkor than in well-nourished children (33) will result in an increased influx of free FA into the liver. Furthermore, decreased peroxisomal ß-oxidation of FA by the liver has been proposed as one of the mechanisms of the pathogenesis of fatty liver, on the basis of the markedly lower concentration of peroxisomes in the liver of severely malnourished children at autopsy than in the livers of children who had recovered from malnutrition (7). It is therefore possible that the fatty liver of the malnourished child can result from an increased availability of FA for hepatic reesterification to triacylglycerol and not from impaired VLDL-apo B-100 synthesis.
In conclusion, the findings of this study indicated that synthesis of the apolipoprotein moiety of VLDL in severely malnourished children with more liver fat is not impaired relative to the synthesis in those with less liver fat. In fact, there seems to be a compensatory response as liver fat increases, suggested by the faster rate of synthesis of apo B-100 in the children with more liver fat. This occurs even in the face of the slower rate of whole-body protein turnover that is characteristic of the severely malnourished state (34).
ACKNOWLEDGMENTS
We are grateful to the physicians and nursing staff of the Tropical Metabolism Research Unit for their care of the children and to Hyacinth Gallimore, Bentley Chambers, Sharon Howell, Margaret Frazer, and Melanie Del Rosario for their excellent work and support in the conduct of the studies and analysis of the samples.
AB contributed to all aspects of the production of this manuscript: the design of the study, execution of the experiments, data collection, analysis and interpretation, and writing of the manuscript. MR and TF were involved in the design of the study, analysis and interpretation of the data, and writing of the manuscript. DS shared responsibility with MR for CT scan measurement and was also involved in writing the manuscript. FJ was involved in the design of the study, analysis and interpretation of the data, and writing of the manuscript. None of the authors had any personal or financial conflict of interest.
REFERENCES