A low-protein diet combined with low-dose endotoxin leads to changes in glucose homeostasis in weanling rats
Abstract
Severe malnutrition is a leading cause of global childhood mortality, and infection and hypoglycemia or hyperglycemia are commonly present. The etiology behind the changes in glucose homeostasis is poorly understood. Here, we generated an animal model of severe malnutrition with and without low-grade inflammation to investigate the effects on glucose homeostasis. Immediately after weaning, rats were fed diets containing 5 [low-protein diet (LP)] or 20% protein [control diet (CTRL)], with or without repeated low-dose intraperitoneal lipopolysaccharide (LPS; 2 mg/kg), to mimic inflammation resulting from infections. After 4 wk on the diets, hyperglycemic clamps or euglycemic hyperinsulinemic clamps were performed with infusion of [U-13C6]glucose and [2-13C]glycerol to assess insulin secretion, action, and hepatic glucose metabolism. In separate studies, pancreatic islets were isolated for further analyses of insulin secretion and islet morphometry. Glucose clearance was reduced significantly by LP feeding alone (16%) and by LP feeding with LPS administration (43.8%) compared with control during the hyperglycemic clamps. This was associated with a strongly reduced insulin secretion in LP-fed rats in vivo as well as ex vivo in islets but signficantly enhanced whole body insulin sensitivity. Gluconeogenesis rates were unaffected by LP feeding, but glycogenolysis was higher after LP feeding. A protein-deficient diet in young rats leads to a susceptibility to low-dose endotoxin-induced impairment in glucose clearance with a decrease in the islet insulin secretory pathway. A protein-deficient diet is associated with enhanced peripheral insulin sensitivity but impaired insulin-mediated suppression of hepatic glycogenolysis.
malnutrition is still a major health problem facing children in the developing world today, being responsible for around 45% of all childhood deaths (11). The severest form of malnutrition is associated, apart from a profound effect on body weight, with metabolic disorders such as hepatic steatosis, both hypoglycemia and hyperglycemia, and oxidative stress (18, 22, 28, 30). Treatment of these children is often challenging, with high inpatient mortality rates (5, 20, 29).
The disturbances in metabolic homeostasis are poorly understood. There are indications that severe malnutrition can lead to impaired glucose tolerance (7). We and others have shown that the impairment in glucose tolerance is in part related to altered pancreatic β-cell function in severely malnourished children (14, 21, 39). There is a paucity of data available on alterations in insulin sensitivity and hepatic glucose metabolism in early postnatal malnutrition. Two studies have reported some indirect evidence of insulin resistance in children and adults with severe malnutrition (3, 37). However, our data based on oral glucose tolerance tests suggested that the whole body insulin sensitivity might not be impaired in severely malnourished children (39). Limited data have suggested improved insulin sensitivity in a low-protein diet-induced rodent model of malnutrition (35, 38).
Many children with severe malnutrition suffer from comorbidities, with one of the most important being infections. Infections in malnourished children are often associated with subtle clinical signs such as persistent low-grade fevers, mild chronic cough, or diarrhea but are associated with undesirable clinical outcomes and a high fatality rate (2, 6, 29). This association was illustrated by a recent randomized clinical trial showing that in presumed uncomplicated children with severe malnutrition, survival was improved significantly by treatment with antibiotic therapy (44).
Infections in severely malnourished children are associated with oxidative stress (18, 30). Infections and oxidative stress have been found to affect pancreatic β-cell function and insulin sensitivity (36, 47, 50). There are currently no data on the effects of low-grade infections on glucose homeostasis in malnutrition. The current study was designed to evaluate the effect of a low-protein diet immediately after weaning as a model of early postnatal malnutrition on insulin secretion and insulin action, the main regulators of glucose homeostasis. We further evaluated the role of repeated low-dose LPS administration as a model of a chronic low-grade infection on the metabolic disturbances in malnourished rats.
MATERIALS AND METHODS
Animals and diet.
Pregnant Wistar rats were placed on regular chow. The offspring were kept with the mother until day 21, after which they were separated into two different diet groups: 1) a semisynthetic 5% protein [low-protein (LP)] diet and 2) a semisynthetic control (CTRL; 20% protein) diet (Harlan, Madison, WI). The protein in the LP diet was replaced by starch, whereas all other components were kept constant. Since in preliminary studies LP-fed animals had reduced food intake (30% less than animals on the CTRL diet), the rats of the CTRL group were pair-fed the same amount as the rats on the LP diet. To exclude metabolic effects of a limited reduction in food intake, we initially added an additional control group of animals fed a regular chow diet ad libitum (CTRL ad libitum). The exact diet composition is provided in Table 1. Animals were kept on the diet for 4 wk. Two weeks into the diet, subgroups of rats on the CTRL and LP diets received an intraperitoneal injection of lipopolysaccharide (2 mg/kg, L2630; Sigma-Aldrich, St. Louis, MO) twice/wk (every 3–4 days). Rats were closely monitored for signs of illness, and weight was measured twice/wk.
| CTRL | |||
|---|---|---|---|
| Component | LP (5% Protein) | 20% Protein | Rregular chow |
| Protein, g/kg | 46 | 207 | 240 |
| Methionine, g/kg | 2.7 | 2.7 | 3.6 |
| Starch, g/kg | 392 | 251 | 383 |
| Sucrose, g/kg | 350 | 350 | NA |
| Maltodextrin, g/kg | 50 | 50 | NA |
| Fat, g/kg | 53 | 53 | 55 |
| Energy, kcal/g | 3.8 | 3.8 | 3.8 |
Surgeries.
Between days 21 and 23, a Micro-Renathane catheter was placed in the right jugular vein of the rats under general anesthesia. Rats were allowed 5–7 days to recover. Only rats that recovered uneventfully and lost <5% of their preoperative body weight were used for the clamp experiments.
Hyperglycemic clamp experiments.
Rats were fasted overnight for 16 h, after which they received a bolus of a 37.5% glucose solution (750 mg/kg) through the right jugular vein. This was immediately followed by a continuous glucose infusion, which was started at 37.5 mg·kg−1·min−1 for 15 min and thereafter titrated to keep glucose concentrations at 17 mmol/l for a total duration of 90 min. Blood glucose measurements were performed by tail sampling every 5–10 min using a glucometer (One Touch Ultra 2; Lifescan, Milpitas, CA). Additional small blood samples were obtained at t = 0, 2, 5, 10, 15, 30, 60, 80, and 90 min for measurement of insulin and C-peptide concentrations via enzyme-linked immunosorbent assays (Biovendor, Asheville, NC). After the clamp, animals were euthanized by a lethal dose of phenobarbital, after which blood and tissues were harvested. Six rats per group were used.
Euglycemic hyperinsulinemic clamp experiments.
Animals were fasted overnight for 16 h, after which a baseline blood sample was taken from the tail. An infusion of [U-13C6]glucose and [2-13C]glycerol was started through the right jugular vein at a rate of 60 and 400 μmol·kg−1·h−1, respectively, and continued for 2.5 h in awake and unrestrained animals for calculation of gluconeogenesis, glycogenolysis, and hepatic glucose production (sum of the two), as described previously (26). Multiple blood samples were taken by tail sampling throughout the infusion period. Following the 2.5-h basal period, a euglycemic hyperinsulinemic clamp was performed. An infusion of insulin (220 mU·kg−1·h−1) and somatostatin (180 μg·kg−1·h−1) was started, dissolved in a solution containing albumin (15 mg/ml), and maintained for 2.5 h based on our work (19, 26) and described by others (13). Concomitantly, an infusion of glucose containing 3% [U-13C6]glucose was started at a rate of 8.06 mmol·kg−1·h−1, whereas the infusion of [U-13C6]glucose alone was stopped and the infusion of [2-13C]glycerol was decreased to a rate of 200 μmol·kg−1·h−1. An adjustable second glucose infusion containing 3% [U-13C6]glucose was started to maintain glucose concentrations between 4.5 and 6.5 mmol/l. Glucose concentrations were measured every 10–15 min using a glucometer (One Touch Ultra; Lifescan). All infusions were continued for 2.5 h. Multiple blood spot samples were taken for the last 60 min of the infusion period. A total of four to six rats per group were used.
Calculations.
Procedures for extraction of glucose from blood spots, derivatization of the extracted glucose, and gas chromatography-mass sprectrometry (GC-MS) measurements of derivatives were all well established and performed as described previously (8, 45, 46). The measured fractional distribution measured by GC-MS (M0–M6) was corrected for natural abundance of 13C by multiple linear regressions, as described by Lee et al. (27) to obtain the excess mole fraction of mass isotopomers due to incorporation of infused labeled compounds, i.e., [2-13C]glycerol and [U-13C6]glucose. Rate of glucose disposal, hepatic gluconeogenesis, and glycogenolysis were calculated as described before (46). The rate of endogenous production of plasma glucose [Ra(glc;endo)] was calculated as Ra(glc;endo) = [MPE(glc;M6)infusate/MPE(glc;M6)plasma − 1] × infusion(glc;M6l), in which MPE(glc;M6) infusate is the mole percent enrichment of infused [U-13C]glucose, MPE(glc;M6) plasma is the mole percent enrichment of plasma [U-13C]glucose, and infusion(glc;M6) is the infusion rate of uniformly labeled [U-13C]glucose. Precursor enrichments [p(glc)] and fractional rate of gluconeogenesis [f(glc)] were calculated using mass isotopomer distribution analyses of glucose derivatives. The absolute gluconeogenic fluxes into plasma glucose [GNG(glc)] were calculated according to GNG(glc) = f(glc) × [total Ra(glc;endo) + infusion(glc;total)].
Assessment of insulin secretion ex vivo.
Pancreatic islets were isolated from overnight-fasted rats using the Ficoll/Histopaque method, as described previously (33). The assessment of insulin secretion and islet insulin content has been described by Koulajian etal. (23). Briefly, isolated islets of rats were preincubated for 1 h at 37°C in KRBH supplemented with 2.8 mmol/l glucose. Thereafter, five rat islets of approximately the same size were incubated in triplicate at 6.5 and 22 mmol/l glucose for 2 h at 37°C. Insulin concentration in the medium was analyzed with an ELISA kit (Merck, Darmstadt, Germany). Insulin concentration detection limits were between 0.5 and 100 ng/ml. The islets used for secretion were then subjected to acid ethanol extraction for measurement of insulin content. Islet reactive oxygen species (ROS) were measured using dihydrodichlorofluorescein diacetate (Life Technologies, Burlington, ON, Canada) followed by fluorescent microscopy (Olympus fluorescence BX51W1 microscope). The collected data were analyzed using the ImageMaster 3 software (43).
Electron microscopy.
Pancreas was harvested immediately following euthanasia and fixed in 2.5% glutaraldehyde for electron microscopy (EM). For EM, tissues were postfixed in OsO4 and embedded in Embed812 (SPI Supplies, West Chester, PA). Ultrathin sections were then prepared, mounted on grids, and stained with uranyl acetate and lead citrate. Images of β-cells from the islets were then captured with a charge-coupled device camera (AMT, Danvers, MA) and an electron microscope (JEM 1011; JEOL USA, Peabody, MA). At least 25 cells were examined from two animals from each group. The ratio of dark (more mature) to light granules was then determined from each cell for each group.
Statistical analysis.
Values reported are means ± SD except for the EM morphometrics, islet ROS data, which are reported as means ± SE. Significance of metabolite and activity levels and fluxes was determined using two-way ANOVA with repeated measures where appropriate by SPSS Statistics package 21.0.0 (IBM SPSS, Armonk, NY). Differences were considered significant at P < 0.05.
RESULTS
Effect of LP diet and LPS on basic characteristics.
As expected, weanling animals fed an LP diet had drastically reduced body weight compared with CTRL diet pair-fed animals (average 67 vs. 201 g; see Table 2). Interestingly, protein restriction led to a decrease in food intake, i.e., the generation a model of a spontaneous hypocaloric intake that, to a lesser extent, also affected the pair-fed control animals in terms of body growth. The caloric intake was 40 ± 8, 39 ± 2, and 63 ± 12 kcal/day in the LP-fed, pair-fed CTRL, and ad libitum CTRL animals. The body weight of the CTRL and ad libitum group was 28% higher than the CTRL pair-fed diet group, and the difference between CTRL and other groups was even greater. Serum albumin concentration, a biochemical indicator of malnutrition, was severely reduced in LP-fed groups compared with CTRL pair-fed groups (average 26 vs. 39 g/l, P < 0.001). LPS administration was associated with reduced activity and food intake for ∼48 h in both groups, although the effect appeared more pronounced in the LP-fed animals. There was no effect of LPS on serum albumin concentrations. Glucose and insulin concentrations after a 16-h fast were not significantly different (Table 2).
| CTRL Ad Libitum | CTRL | CTRL + LPS | LP | LP + LPS | |
|---|---|---|---|---|---|
| Body weight, g | 258 ± 37* | 201 ± 11 | 209 ± 27 | 67 ± 11* | 64 ± 12* |
| Albumin, g/l | 38 ± 3 | 39 ± 26 | 39 ± 3 | 26 ± 2* | 27 ± 5* |
| Glucose, mmol/l | 5.5 ± 1.1 | 6.2 ± 0.8 | 6.5 ± 1.2 | 5.0 ± 1.5 | 5.5 ± 0.6 |
| Insulin, ng/ml | 0.3 ± 0.3 | 0.8 ± 0.7 | 0.3 ± 0.1 | 0.6 ± 0.5 | 0.4 ± 0.3 |
Insulin secretion assessed in response to a hyperglycemic clamp.
The results of the hyperglycemic clamp are shown in Fig. 1. The glucose concentration curves are shown in Fig. 1A. There was a lower glucose concentration after LP feeding compared with CTRL feeding (P < 0.01) during the first 15 min of the clamp, when all animals received a similar glucose infusion rate per body weight. Glucose concentrations at steady state, i.e., during the last 20 min of the clamp, were similar in all groups. The average glucose infusion rate (Ginf) of the last 20 min of the clamp was decreased 16% by the LP diet compared with the CTRL diet (P < 0.05) and by LPS administration with 20.5% reduction in CTRL group and 33.1% in the LP diet group, consistent with reduced glucose clearance (Fig. 1B). The acute insulin secretory response during the first 15 min of the hyperglycemic clamp was blunted (P < 0.005) by LP feeding compared with CTRL feeding without an effect of LPS administration (Fig. 1C). Similarly, there was a lower steady-state insulin concentration in the LP-fed groups compared with CTRL fed animals (P < 0.005; Fig. 1C). Similarly, there was a significantly lower steady-state insulin concentration in the LP-fed groups (2.3 ± 0.5 ng/ml) compared with CTRL pair-fed animals (3.9 ± 0.7 ng/ml) (P < 0.005) (Fig. 1C). There was no effect of LPS administration on insulin secretion. Results were similar, albeit less pronounced, for C-peptide concentrations (Fig. 1D). C-peptide concentrations in LP-fed animals were lower compared with CTRL-fed animals during the first 15 min of the clamp (P < 0.05). C-peptide concentrations during steady state were also lower in LP-fed animals (P < 0.05) without an effect of LPS. There was no effect of CTRL ad libitum feeding compared with CTRL pair-feeding for any of the parameters tested.
Fig. 1.Plasma glucose (A), insulin (C), and C-peptide (D) concentrations during the hyperglycemic clamp. Glucose infusion rate (Ginf; B) during the last 20 min of the hyperglycemic clamp; n = 6 in all groups. Black circles and long black dashed line, control (CTRL) ad libitum; red diamonds and red solid line, CTRL; green circles and short green dashed line, CTRL + LPS; purple triangles and short purple dashed line, low-protein (LP) diet; blue squares and blue dotted line, LP diet + LPS. Light gray bar, CTRL ad libitum; open bar, CTRL; black bar, CTRL + LPS; dark gray bar, LP diet; hatched bar, LP diet + LPS. *P < 0.01, LP vs. CTRL; †P < 0.005, LP vs. CTRL; ‡P < 0.05, LP vs. CTRL; §P < 0.05, LPS vs. no LPS. Data were analyzed by 2-way repeated-measures ANOVA.
Glucose-stimulated insulin secretion and islet insulin content in isolated islets.
Figure 2 shows the insulin response in isolated islets from rats in four different groups. Although the islet insulin content appeared to be lower after LP feeding, it did not reach statistical significance (P = 0.08; Fig. 2A). At 6.5 mmol/l glucose, there was an 84.1% reduction in insulin secretion after LP feeding, with no effect of LPS administration (P < 0.001; Fig. 2B). At 22 mmol/l glucose, insulin secretion was reduced by LP feeding by 50.5%, with a trend for even more reduced secretion after LPS administration (P < 0.005; Fig. 2B). Pancreatic β-cells have low antioxidant defences and are susceptible to decreased function induced by ROS, and therefore, we determined the effect of LP feeding and LPS on ROS production. There was no significant effect of LP feeding or LPS on ROS concentrations (P = 0.1; Fig. 2C).
Fig. 2.Effect of LP diet and LPS on islet insulin content (A) and insulin secretory response (B) to glucose and islet reactive oxygen species (ROS) levels (C); n = 6/group. Open bars, CTRL diet; black bars, CTRL + LPS; dark gray bars, LP diet; hatched bars, LP diet + LPS. *P < 0.001, LP vs. CTRL; †P < 0.005, LP vs. CTRL using a 2-way ANOVA. In C, data are expressed as %saline ± SE.
Electron microscopy.
Light microscopy with insulin staining did not reveal major differences related to diet. In particular, there was no difference in the relative surface area of the islets to pancreas (data not shown). There was a lower dark/light granule ratio (index of granule maturity) in the LP-fed groups compared with the CTRL groups (Fig. 3). Administration of low-dose LPS did not have any effect on this ratio in either group of animals.
Fig. 3.Electron micrographs of pancreatic islets after 4 wk of CTRL or LP diet. A, images A1 and A3: electron microscopty (EM) of islet β-cell from a CTRL diet-fed animal. Note the dominance of dark secretory granules in the cytoplasm. Images A2 and A4: EM from LP-fed animal is shown in image A4. Note the dominance of the light secretory granules. B: morphometric analyses of islet EM measured by dark granule/light granule density ratio in different groups. A minimum of 5 slides/group were analyzed from 3 animals to calculate islets/pancreatic area. A minimum of 25 cells/group were analyzed from 2 animals to calculate dark granule/light granule density ratio. Data are expressed as a means ± SE. *P < 0.05, LP vs. CTRL.
Whole body insulin sensitivity, gluconeogenesis, and glycogenolysis.
To evaluate glucose metabolism and insulin sensitivity, we performed stable isotope experiments during the basal state and during a euglycemic hyperinsulinemic clamp. We evaluated the M6 enrichments to evaluate whether we had achieved an isotopic steady state. M6 enrichments were similar with 15.2 and 15.7% at 30 min before and at the end of the basal period, respectively. For the clamp period, enrichments were clearly at isotopic steady state with 18.9 and 19.0% at 30 min before and at the end of the clamp period, respectively. Ginf and glucose concentrations were similar during the last 15 min of the clamp, which could suggest that animals reached a metabolic steady state. Insulin concentrations were 9.8 ± 0.2 and 10.1 ± 0.3 ng/ml in the LP-fed and control animals, respectively. Free fatty acid concentrations were 0.20 ± 0.05 and 0.19 ± 0.13 mmol/l in the LP-fed and control animals, respectively. Ginf (i.e., whole body insulin sensitivity) during the euglycemic clamp were higher in the LP groups (1.60 ± 0.12 mmol·kg−1·min−1) compared with the CTRL groups (0.95 ± 0.05 mmol·kg−1·min−1) (P < 0.001) (Fig. 4A). Ginf rates were unaffected by LPS administration. Basal rate of disappearance of glucose, which is equal to rate of appearance, i.e., endogenous glucose production (RdGlc), was similar in all groups (Fig. 4B). However, during the clamp period, RdGlc was elevated in LP-fed animals (2.0 ± 0.13 mmol·kg−1·min−1) compared with control animals (1.2 ± 0.06 mmol·kg−1·min−1) (P < 0.001), indicating enhanced peripheral insulin sensitivity. Basal gluconeogenesis rates were similar in the different groups, and hepatic gluconeogenesis was inhibited during the clamp to a similar degree in all groups (Fig. 4C). However, basal hepatic glycogenolysis rates were borderline higher in LP-fed animals (209.6 ± 16.9 μmol·kg−1·min−1) vs. control (138.7 ± 8.3 μmol·kg−1·min−1) (P = 0.07) and significantly higher during the steady-state clamp (333.1 ± 20.5 μmol·kg−1·min−1) compared with the CTRL groups (175.2 ± 14.1 μmol·kg−1·min−1) (P < 0.001; Fig. 4D). There was no effect of LPS administration on glycogenolysis rates.
Fig. 4.Ginf (A), glucose rate of disappearance (RdGlc; B), gluconeogenesis (GNG; C), and glycogenolysis (D) during the basal and clamp period; n = 4–6/group. Open bars, CTRL; black bars, CTRL + LPS; dark gray bars, LP diet; hatched bars, LP diet + LPS. *P < 0.001, LP vs. CTRL using a 2-way repeated-measures ANOVA.
DISCUSSION
Despite the consensus that severe malnutrition is still a widespread condition that is associated with a high mortality rate, few studies have focused on understanding the pathophysiological mechanisms that lead to disordered metabolic homeostasis in this condition. In the current project, we focused on studying hepatic and peripheral glucose metabolism and pancreatic islet function and morphology in a novel in vivo rat model of early postnatal severe malnutrition with or without chronic low-grade inflammation.
As expected, the low-protein dietary intervention led to a loss of growth, which has been reported previously (16, 42). This is consistent with what is seen in children with severe malnutrition, which is defined by a weight-for-height z-score of less than −3 (49). The main results indicate that 1) there is decreased responsiveness of pancreatic β-cells that is related to decreased basal and glucose-stimulated insulin secretion, with relatively preserved islet insulin content after LP feeding; and 2) severe malnutrition is associated with enhanced peripheral insulin sensitivity, appropriately suppressed hepatic gluconeogenesis but increased hepatic glycogenolysis, as evidenced during the euglycemic hyperinsulinemic clamp.
There have been a number of clinical reports indicating changes in pancreatic β-cell function in malnourished patients (14, 21, 39). Pancreatic function in these children was generally assessed using oral glucose tolerance tests. However, we recently reported severely impaired glucose absorption in malnourished children (9). Decreased absorption of glucose would ultimately also influence insulin secretion, making the use of an oral glucose tolerance test less reliable in malnourished children. There are also animal studies focusing on pancreatic β-cell function during malnutrition. Although some data have indicated a normal insulin response after a low-protein diet (16), most in vivo studies using oral or intravenous glucose tolerance tests (32, 35, 42) and isolated pancreatic islet (17, 38) data have shown decreased pancreatic β-cell response. In this study, we provide further in vivo and in vitro evidence for a decrease in pancreatic β-cell insulin secretion by using the gold standard for assessing pancreatic function, i.e., the hyperglycemic clamp technique.
Our data indicate that the reduced insulin secretory response is not related to a significant lack of insulin, as islet density and insulin content were not significantly reduced in the LP-fed animals. Our electron microscopy did reveal an increased presence of pale islet granules, probably representing an immature population of progranules (34). This is consistent with data obtained from postmortem-acquired pancreatic tissue of severely malnourished children (12). Only a handful of studies have been published that examined the pathophysiological mechanisms responsible for pancreatic β-cell failure in malnutrition. Reduced insulin secretion in protein-malnourished mice might be related to alterations in the stimulus-secretion coupling of pancreatic β-cells, including more hyperpolarized membrane potential attributed to higher activity of ATP-sensitive potassium ion channels, thereby leading to reduced Ca2+ influx and an inhibition of insulin release (38). Limited animal data also suggested changes in the cAMP/protein kinase A regulation of the insulin secretion pathway (17) and possibly changes in insulin receptor and insulin receptor substrate activity (35).
In this study, we did observe an additional effect of LPS in the LP-fed group on glucose clearance during the hyperglycemic clamp. However, we could not detect a statistically significant effect of LPS on insulin secretion. LPS has been shown to rapidly decrease insulin secretion in isolated islets ex vivo (48). More recent data generated evidence that the effect of LPS is mediated through Toll-like receptor 4 (4). In addition, we and others have shown that impaired insulin secretion can be mediated through oxidative stress (24, 25, 33, 50). Pancreatic β-cells are relatively vulnerable to oxidative stress due to their relatively low free radical detoxifying and redox-regulating enzymes (31). Impairment of pancreatic β-cell function is ameliorated or prevented by antioxidant therapy in animal models (33) and in humans (51). In children, severe malnutrition is associated with oxidative stress (18, 30). However, although there was a trend for increased islet ROS levels especially in the LP group with LPS, we could not find a significant effect of either LP feeding or LPS. Although the dose of LPS used in this study did affect glucose clearance, it is likely that the dose was not sufficiently high to strongly affect insulin secretion. The dose we used for this study was aimed at trying to mimick a low-grade chronic inflammation that is often clinically observed in malnourished children, e.g., related to tuberculosis, chronic intestinal infection. We intentionally did not want to create an overwhelming condition of sepsis that would also potentially mask any metabolic effects caused by malnutrition. It would have been interesting to evaluate different doses of LPS, but unfortunately, this was not feasible due to the labor-intensive property of the clamp procedures.
Whole body insulin sensitivity was found to be higher in LP-fed rats compared with regularly fed rats. Our study is the first to study insulin sensitivity using the gold standard, i.e., the euglycemic hyperinsulinemic clamp technique in malnutrition. Reis et al. (35) found normal glucose response during a glucose tolerance test despite decreased insulin response in malnourished rats, suggesting enhanced insulin action. Other animal and human studies using glucose tolerance tests also found some evidence for enhanced insulin sensitivity in malnutrition (3, 16, 39). The increased insulin sensitivity may be due to lower body weight and fat mass and/or may be adaptive to low insulin secretion. Intriguingly, repeated low-dose LPS administration had no effect on insulin sensitivity in our experiments. Sepsis and administration of LPS have been shown to induce insulin resistance both in animal studies as well as in human subjects (1, 47).
In this study, we did find evidence for some degree of hepatic insulin resistance. Although gluconeogenesis was similarly suppressed during the euglycemic clamp in both control and LP-fed rats, we observed increased rates of glycogenolysis during the euglycemic clamp in the LP-fed rats. As far as we know, no studies that have specifically addressed the regulation of hepatic glycogenolysis in childhood malnutrition exist. In conditions of insulin resistance, glycogenolysis and gluconeogenesis are usually suppressed to a similar extent (10). Although we are unable to confirm this, we cannot exlude that differences in corticosterone or catecholamines levels contributed to the different glycogenolysis rates between the groups.
There are several limitations to this study. First, our sample size was relatively small. It could be that the study was underpowered to detect between-group changes in islet insulin concentration and basal glycogenolysis. However, these values were physiologically close to each other anyway, so they likely would not have affected our main conclusions. In addition, although dietary calorie intake was similar among the animals, weight changes among the groups increased over time, leading to a higher caloric intake in the LP-fed animals when expressed per kilogram body weight. All kinetic analyses were expressed per kilogram body weight. If expressed per lean body weight, it could have influenced values for glucose infusion rates. However, we felt that, at least when related to hepatic glucose metabolism, expression per total body weight was acceptable, as liver weight per body weight was similar in all groups. Unfortunately, we did not have the possibility to determine body composition in our animals. Furthermore, the insulin dose used in the euglycemic hyperinsulinemic clamp study was relatively modest. A higher or loading dose of insulin would have likely led to a suppression of hepatic glycogenolysis in the control group but could have prevented us from finding a difference in glycogenolysis between the low-protein and control groups. The absence of an insulin-loading dose could very well have prevented us from reaching a complete metabolic steady state, which would affect glucose clearance rates and possibly serum FFA concentrations. However, we are confident that this would not have affected our conclusion on the difference found between the groups. We tried extensively to insert cannulas into carotid arteries, which would allow serial sampling of larger volumes for analysis of insulin concentrations to better determine whether a steady state was reached. We were able to insert them successfully, but we found that for the small malnourished animals we had to use such small diameter cannulas that it was not feasible to withdraw blood from them without clotting the lines. Tail sampling is more stressful for rats, especially if one needs to collect larger volumes of blood and manipulate the tail longer. For ethical reasons and because of the effect of stress on glucose homeostasis, we decided to take only small bloodspot samples for glucose and enrichment measurements and have one final blood draw at the time of euthanasia, which showed similar insulin concentrations in control and LP-fed animals.
In conclusion, our study shows that reduced insulin secretory response during severe malnutrition is compensated at least in part by increased peripheral insulin sensitivity. In addition, low-protein diet-induced malnutrition leads to an enhanced rate of glycogenolysis. The results of this study might explain the pathophysiology of hyperglycemia in children with severe malnutrition. Improving glucose tolerance in sick, severely malnourished patients could potentially lead to a decrease in mortality. However, in low-resource settings, interventions with insulin might not be feasible and potentially high risk due to the increased insulin sensitivity of malnutrition. Since impaired β-cell function in the context of refeeding could lead to hyperglycemia, whereas sudden improvements in insulin release with enhanced insulin sensitivity during refeeding may cause hypoglycemia, further studies are also required in a refeeding model. These results provide insight into the metabolic disturbances occurring during early postnatal malnutrition in vivo and could have important consequences for treatment of severely malnourished children.
GRANTS
G. F. Lewis holds the Sun Life Financial Chair in Diabetes and the Drucker Family Chair in Diabetes Research. R. H. J. Bandsma is supported through a new investigator grant from the
DISCLOSURES
All authors report that there is no duality of interest related their contributions to this article.
AUTHOR CONTRIBUTIONS
R.H.B., C.A., T.H.v.D., and G.F.L. conception and design of research; R.H.B., C.A., K.K., T.v.Z., and C.X. performed experiments; R.H.B., C.A., K.K., L.Z., T.v.Z., and G.F.L. analyzed data; R.H.B., C.A., K.K., L.Z., T.v.Z., T.H.v.D., C.X., A.G., and G.F.L. interpreted results of experiments; R.H.B., C.A., K.K., L.Z., T.v.Z., T.H.v.D., and C.X. prepared figures; R.H.B., L.Z., A.G., and G.F.L. drafted manuscript; R.H.B., L.Z., C.X., A.G., and G.F.L. edited and revised manuscript; R.H.B. approved final version of manuscript.
ACKNOWLEDGMENTS
We thank Roberta Allgayer de Moraes (University Medical Centre Groningen, Groningen, The Netherlands) for excellent technical assistance.
REFERENCES
- 1. . Insulin resistance and substrate utilization in human endotoxemia. J Clin Endocrinol Metab 85: 3770–3778, 2000.
Crossref | PubMed | Web of Science | Google Scholar - 2. . Mortality in severely malnourished children with diarrhoea and use of a standardised management protocol. Lancet 353: 1919–1922, 1999.
Crossref | PubMed | Web of Science | Google Scholar - 3. . Glucose tolerance and insulin sensitivity in malnourished children. Br J Nutr 27: 585–592, 1972.
Crossref | PubMed | Web of Science | Google Scholar - 4. . Lipopolysaccharides impair insulin gene expression in isolated islets of Langerhans via Toll-Like Receptor-4 and NF-κB signalling. PLoS One 7: e36200, 2012.
Crossref | PubMed | Web of Science | Google Scholar - 5. . WHO guidelines for management of severe malnutrition in rural South African hospitals: effect on case fatality and the influence of operational factors. Lancet 363: 1110–1115, 2004.
Crossref | PubMed | Web of Science | Google Scholar - 6. . Bacteraemia in severely malnourished children in an HIV-endemic setting. Ann Trop Paediatr 26: 319–328, 2006.
Crossref | PubMed | Google Scholar - 7. . Carbohydrate metabolism in kwashiorkor. Lancet 2: 662–665, 1965.
Crossref | PubMed | Web of Science | Google Scholar - 8. . Hepatic de novo synthesis of glucose-6-phosphate is not affected in PPAR-deficient mice but is preferentially directed towards hepatic glycogen stores after a short-term fast. J Biol Chem 279: 8930–8937, 2004.
Crossref | PubMed | Web of Science | Google Scholar - 9. . Impaired glucose absorption in children with severe malnutrition. J Pediatr 158: 282–287.e1 2011.
Crossref | PubMed | Web of Science | Google Scholar - 10. . Obesity and type 2 diabetes impair insulin-induced suppression of glycogenolysis as well as gluconeogenesis. Diabetes 54: 1942–1948, 2005.
Crossref | PubMed | Web of Science | Google Scholar - 11. ;
Maternal and Child Nutrition Study Group . Maternal and child undernutrition and overweight in low-income and middle-income countries. Lancet 382: 427–451, 2013.
Crossref | PubMed | Web of Science | Google Scholar - 12. . Ultrastructure of the islets of Langerhans in protein-energy malnutrition. West Indian Med J 42: 101–106, 1993.
PubMed | Web of Science | Google Scholar - 13. . Glucose transporter-1 in the hypothalamic glial cells mediates glucose sensing to regulate glucose production in vivo. Diabetes 60: 1901–1906, 2011.
Crossref | PubMed | Web of Science | Google Scholar - 14. . Blood sugar and serum insulin response in protein-energy malnutrition. J Trop Pediatr 44: 139–141, 1998.
Crossref | PubMed | Web of Science | Google Scholar - 15. . Blood glucose concentration and outcome of critical illness: the impact of diabetes. Crit Care Med 36: 2249–2255, 2008.
Crossref | PubMed | Web of Science | Google Scholar - 16. . Increased insulin action in the rat after protein malnutrition early in life. Diabetologia 34: 559–564, 1991.
Crossref | PubMed | Web of Science | Google Scholar - 17. . Decreased insulin secretion in islets from rats fed a low protein diet is associated with a reduced PKAalpha expression. J Nutr 134: 63–67, 2004.
Crossref | PubMed | Web of Science | Google Scholar - 18. . Free radicals in the pathogenesis of kwashiorkor. Proc Nutr Soc 46: 53–68, 1987.
Crossref | PubMed | Web of Science | Google Scholar - 19. . Differential effects of pharmacological liver X receptor activation on hepatic and peripheral insulin sensitivity in lean and ob/ob mice. Am J Physiol Endocrinol Metab 289: E829–E838, 2005.
Link | Web of Science | Google Scholar - 20. . How can we improve the care of severely malnourished children in Africa? PLoS Med 4: e45, 2007.
Crossref | PubMed | Web of Science | Google Scholar - 21. . Insulin secretion in kwashiorkor. J Clin Endocrinol Metab 35: 63–66, 1972.
Crossref | PubMed | Web of Science | Google Scholar - 22. . Hypoglycaemia in infantile malnutrition. Acta Paediatr Scand Suppl 172: 119+, 1967.
Crossref | Google Scholar - 23. . NADPH oxidase inhibition prevents beta cell dysfunction induced by prolonged elevation of oleate in rodents. Diabetologia 56: 1078–1087, 2013.
Crossref | PubMed | Web of Science | Google Scholar - 24. . NADPH oxidase inhibition prevents beta cell dysfunction induced by prolonged elevation of oleate in rodents. Diabetologia 56: 1078–1087, 2013.
Crossref | PubMed | Web of Science | Google Scholar - 25. . Overexpression of glutathione peroxidase 4 prevents β-cell dysfunction induced by prolonged elevation of lipids in vivo. Am J Physiol Endocrinol Metab 305: E254–E262, 2013.
Link | Web of Science | Google Scholar - 26. . Chronic prednisolone treatment aggravates hyperglycemia in mice fed a high-fat diet but does not worsen dietary fat-induced insulin resistance. Endocrinology 153: 3713–3723, 2012.
Crossref | PubMed | Web of Science | Google Scholar - 27. . Mass isotopomer analysis: theoretical and practical considerations. Biol Mass Spectrom 20: 451–458, 1991.
Crossref | PubMed | Google Scholar - 28. . Malnutrition may worsen the prognosis of critically ill children with hyperglycemia and hypoglycemia. JPEN J Parenter Enteral Nutr 37: 335–341, 2013.
Crossref | PubMed | Web of Science | Google Scholar - 29. . Children with severe malnutrition: can those at highest risk of death be identified with the WHO protocol? PLoS Med 3: e500, 2006.
Crossref | PubMed | Web of Science | Google Scholar - 30. . Increased oxidative stress in kwashiorkor. J Pediatr 137: 421–424, 2000.
Crossref | PubMed | Web of Science | Google Scholar - 31. . Diabetes associated cell stress and dysfunction: role of mitochondrial and non-mitochondrial ROS production and activity. J Physiol 583: 9–24, 2007.
Crossref | PubMed | Web of Science | Google Scholar - 32. . Repercussions of chronic protein-calorie malnutrition on glucose homeostasis in the rat. Diabetologia 30: 946–951, 1987.
Crossref | PubMed | Web of Science | Google Scholar - 33. . Free fatty acid-induced reduction in glucose-stimulated insulin secretion: evidence for a role of oxidative stress in vitro and in vivo. Diabetes 56: 2927–2937, 2007.
Crossref | PubMed | Web of Science | Google Scholar - 34. . Direct identification of prohormone conversion site in insulin-secreting cells. Cell 42: 671–681, 1985.
Crossref | PubMed | Web of Science | Google Scholar - 35. . Glucose-induced insulin secretion is impaired and insulin-induced phosphorylation of the insulin receptor and insulin receptor substrate-1 are increased in protein-deficient rats. J Nutr 127: 403–410, 1997.
Crossref | PubMed | Web of Science | Google Scholar - 36. . Beta-cell uncoupling protein 2 regulates reactive oxygen species production, which influences both insulin and glucagon secretion. Diabetes 60: 2710–2719, 2011.
Crossref | PubMed | Web of Science | Google Scholar - 37. . Insulin secretion and glucose tolerance in adults with protein-calorie malnutrition. Metab Clin Exp 24: 1073–1084, 1975.
Crossref | PubMed | Web of Science | Google Scholar - 38. . Reduced insulin secretion in protein malnourished mice is associated with multiple changes in the beta-cell stimulus-secretion coupling. Endocrinology 151: 3543–3554, 2010.
Crossref | PubMed | Web of Science | Google Scholar - 39. . Kwashiorkor and marasmus are both associated with impaired glucose clearance related to pancreatic β-cell dysfunction. Metabolism 61: 1224–1230, 2012.
Crossref | PubMed | Web of Science | Google Scholar - 40. . Diabetes does not alter mortality or hemostatic and inflammatory responses in patients with severe sepsis. Crit Care Med 38: 539–545, 2010.
Crossref | PubMed | Web of Science | Google Scholar - 41. . Persistent reduction of pancreatic beta-cell mass after a limited period of protein-energy malnutrition in the young rat. Diabetologia 35: 939–945, 1992.
Crossref | PubMed | Web of Science | Google Scholar - 42. . Persistent impairment of insulin secretory response to glucose in adult rats after limited period of protein-calorie malnutrition early in life. Diabetes 36: 454–458, 1987.
Crossref | PubMed | Web of Science | Google Scholar - 43. . Evidence for a role of superoxide generation in glucose-induced beta-cell dysfunction in vivo. Diabetes 56: 2722–2731, 2007.
Crossref | PubMed | Web of Science | Google Scholar - 44. . Antibiotics as part of the management of severe acute malnutrition. N Engl J Med 368: 425–435, 2013.
Crossref | PubMed | Web of Science | Google Scholar - 45. . Quantification of hepatic carbohydrate metabolism in conscious mice using serial blood and urine spots. Anal Biochem 322: 1–13, 2003.
Crossref | PubMed | Web of Science | Google Scholar - 46. . Acute inhibition of hepatic glucose-6-phosphatase does not affect gluconeogenesis but directs gluconeogenic flux toward glycogen in fasted rats. A pharmacological study with the chlorogenic acid derivative S4048. J Biol Chem 276: 25727–25735, 2001.
Crossref | PubMed | Web of Science | Google Scholar - 47. . Mechanisms of insulin resistance during acute endotoxemia. Endocrinology 134: 2072–2078, 1994.
Crossref | PubMed | Web of Science | Google Scholar - 48. . Evidence of expression of endotoxin receptors CD14, toll-like receptors TLR4 and TLR2 and associated molecule MD-2 and of sensitivity to endotoxin (LPS) in islet beta cells. Clin Exp Immunol 133: 208–218, 2003.
Crossref | PubMed | Web of Science | Google Scholar - 49.
World Health Organization . Management of Severe Malnutrition: A Manual for Physicians and Other Senior Health Workers. Geneva: World Health Organization, 1999.
Google Scholar - 50. . Oxidative stress is a mediator of glucose toxicity in insulin-secreting pancreatic islet cell lines. J Biol Chem 279: 12126–12134, 2004.
Crossref | PubMed | Web of Science | Google Scholar - 51. . Oral taurine but not N-acetylcysteine ameliorates NEFA-induced impairment in insulin sensitivity and beta cell function in obese and overweight, non-diabetic men. Diabetologia 51: 139–146, 2008.
Crossref | PubMed | Web of Science | Google Scholar
