Glucagon-like peptide-2 induces rapid digestive adaptation following intestinal resection in preterm neonates

Abstract

Short bowel syndrome (SBS) is a frequent complication after intestinal resection in infants suffering from intestinal disease. We tested whether treatment with the intestinotrophic hormone glucagon-like peptide-2 (GLP-2) increases intestinal volume and function in the period immediately following intestinal resection in preterm pigs. Preterm pigs were fed enterally for 48 h before undergoing resection of 50% of the small intestine and establishment of a jejunostomy. Following resection, pigs were maintained on total parenteral nutrition (TPN) without (SBS, n = 8) or with GLP-2 treatment (3.5 μg/kg body wt per h, SBS+GLP-2, n = 7) and compared with a group of unresected preterm pigs (control, n = 5). After 5 days of TPN, all piglets were fed enterally for 24 h, and a nutrient balance study was performed. Intestinal resection was associated with markedly reduced endogenous GLP-2 levels. GLP-2 increased the relative absorption of wet weight (46 vs. 22%), energy (79 vs. 64%), and all macronutrients (all parameters P < 0.05). These findings were supported by a 200% increase in sucrase and maltase activities, a 50% increase in small intestinal epithelial volume (P < 0.05), as well as increased DNA and protein contents and increased total protein synthesis rate in SBS+GLP-2 vs. SBS pigs (+100%, P < 0.05). Following intestinal resection in preterm pigs, GLP-2 induced structural and functional adaptation, resulting in a higher relative absorption of fluid and macronutrients. GLP-2 treatment may be a promising therapy to enhance intestinal adaptation and improve digestive function in preterm infants with jejunostomy following intestinal resection.

short bowel syndrome (SBS) is a condition characterized by malabsorption, diarrhea, dehydration, and malnutrition resulting from excessive loss of absorptive area following intestinal resection. In neonates, intestinal resection can be a consequence of congenital anomalies or necrotizing enterocolitis (NEC), a serious gastrointestinal inflammatory condition that mainly affects preterm infants. NEC is the primary cause of neonatal SBS (17) and occurs in 5–10% of hospitalized preterm infants (46). No effective treatment of neonatal SBS is available, and long-term parenteral nutrition (PN) is often needed (12). However, long-term PN can lead to catheter-related bloodstream infection, thrombosis, intestinal failure-associated liver disease, as well as gastrointestinal complications (15). Furthermore, clinical trials with growth factor peptides, such as growth hormone, have shown limited efficacy in reducing the need for PN (11).

Glucagon-like peptide-2 (GLP-2) is a 33-amino-acid peptide, secreted by the enteroendocrine L cells, which indirectly or directly stimulates intestinal growth and nutrient absorption (7–9, 48). The rapid intestinal growth and maturation in neonates appears to be GLP-2 mediated (32), and exogenous GLP-2 is known to stimulate intestinal brush-border enzyme expression and decrease apoptosis and proteolysis in preterm pigs (4, 31). Furthermore, GLP-2 increases nutrient absorption (37) and stimulates intestinal blood flow in piglets (13, 14).

In adult SBS rats, GLP-2 has been shown to increase bowel weight, villus height, and crypt cell proliferation (27, 28, 41). In adult mice, resection causes expansion of intestinal stem cells (5), and early postsurgical GLP-2 treatment augments this stem cell expansion (10). However, the exact cellular mechanism of action of GLP-2 remains unclear (33). In human adults, GLP-2 improves nutritional status in patients with SBS, both when provided as the native GLP-2 (20) and as a long-acting analog resistant to degradation by dipeptidyl peptidase-4 (DPP-4) (19, 23, 24).

The physiological role and therapeutic potential of GLP-2 for the neonatal SBS gut, particularly after preterm birth, remains to be established. A single study has examined the effect of GLP-2 treatment in a juvenile pig SBS model and reported decreased brush-border enzyme activities, villous atrophy, and decreased weight gain in the GLP-2-treated animals (29). In contrast, increasing plasma GLP-2 levels by supplementing total parenteral nutrition (TPN) with short-chain fatty acids augmented mucosal growth after intestinal resection in neonatal pigs (1). An important aspect of all previously reported animal SBS studies is that they have employed a surgical model with resection of the majority of the small intestine (50–80%), leaving the ileum and colon in continuity (26, 27, 29). This procedure elevates the endogenous GLP-2 release after resection (26, 27, 29), as shown also in adult patients with SBS (21, 24). In a recent study using neonatal pigs, an ileal resection model was described along with the well-established jejunal resection model, but both models still had a colon in continuity (44). Here the ileum was found to be important for intestinal adaptation as well, meaning not just the colon plays a role, as supported by clinical findings (6). However, extensive intestinal resection in neonates most often results in a proximal stoma at least for some weeks or months. The removal of the ileum and colon from the luminal nutrient flow may lead to decreased endogenous GLP-2 release and likely contributes to poor intestinal adaptation in these infants. We hypothesized that a 50% intestinal resection with a jejunostomy in premature piglets would lead to severe compromise of intestinal function and decreased GLP-2 release. Furthermore, we hypothesized that early postsurgical treatment with exogenous GLP-2 treatment would stimulate intestinal functions and adaptation, leading to improved nutrient absorption.

MATERIALS AND METHODS

Animals.

Preterm pigs (Danish Landrace × Yorkshire × Duroc) were obtained from two litters delivered by Cesarean section (2) at 92% of gestation, showing signs of prematurity as previously established (35). When stable respiration was attained, the piglets were transferred to infant incubators. The pigs were fitted with vascular and orogastric catheters and given maternal plasma to ensure immunization (35). The pigs were fed bovine colostrum for 48 h as boluses of 15 ml/kg body wt per 3 h. On postnatal day 2, enteral feeding was stopped 6 h before surgery. TPN was initiated 30 min before anesthesia.

Surgical procedures.

Pigs were stratified into three groups: unresected controls (n = 5), SBS+GLP-2 (n = 7), and SBS+vehicle (n = 8). Pigs were given methadone (0.1 mg/kg im) and intubated, and anesthesia was induced and maintained with inhalation of isofluorane (1–2%, Isoba; Schering-Ploug Animal Health, Ballerup, Denmark). Through a lower left paramedian incision, the ileocecal plicae was located. The ileum was divided, and the distal end was ligated with a 4-0 monocryl (Monocryl, Ethicon; Johnson & Johnson Nordic, Birkerød, Denmark) at the ileocecal junction. The distal small intestine was resected corresponding to 50% of total small intestinal length. Total intestinal length was estimated from a series of previous studies in preterm pigs as 280 cm/(BW^0.60). A 0.5-cm everted stoma was created at the skin level in the left subcostal region with Monocryl 4-0 sutures. The abdominal incision was closed in layers with interrupted 3-0 sutures in the fascia (VICRYL, Ethicon; Johnson & Johnson Nordic), and a continuous nonabsorbable 3-0 suture (Ethilon, Ethicon; Johnson & Johnson Nordic) in the skin. The skin around the stoma was cleaned with acetone for optimal stoma bag attachment, and Eakin wound pouch (839260; Focuscare, Randers, Denmark), was attached. For optimal skin contact and to prevent leakage of stoma fluids, a Hollister Adapt Paste (Dansac & Hollister, Fredensborg, Denmark) was used. The stoma bags were changed when needed and at least every 24 h.

Twelve hours before the surgery and every 12 h for the following 3 days, animals received ampicillin (50 mg/kg body wt; Pentrexyl; Bristol-Myers Sqibb, Bromma, Sweden) and gentamycin (5 mg/kg body wt; Hexamycin; Sandoz A/S, Copenhagen, Denmark), both given intramuscularly. Intramuscular analgesic was given every 12 h following surgery for 3 days (buprenorphine 0.06 mg/kg body wt; Temgesic; Reckitt Benckiser Pharmaceuticals, Slough, United Kingdom). After surgery, piglets were given TPN at a dosage of 4 ml/kg body wt per hour for 24 h and increased daily over the following days by 1 ml/kg body wt per hour. The TPN used was Nutriflex lipid plus (Braun, Melsungen, Germany), adjusted to piglets' needs with added amino acids (Vamin; Fresenius Kabi, Uppsala, Sweden) to provide the following: energy: 300–600 kJ/kg body wt per day; amino acids: 4.3–8.6 g/kg body wt per day; carbohydrate: 6.9–13.8 g/kg body wt per day; lipid: 3.0–6.0 g/kg body wt per day. Human native GLP-2 (1–33) (kindly donated by Dr. Lars Thim, Novo Nordisk, Bagsværd, Denmark) was dissolved in 0.1% human serum albumin and injected into the TPN and given as a continuous infusion of 3.5 μg/kg body wt per hour. In addition, 2 ml/kg body wt per hour of 0.9% NaCl was given intravenously to keep animals hydrated. Piglets were weighed daily and clinically evaluated for signs of general discomfort, infection, dehydration, or edema resulting from excessive hydration.

In vivo galactose test and nutrient balance study.

On day 4 after resection, each pig received 15 ml/kg body wt of a 5% galactose solution in 0.9% saline through the orogastric feeding tube. Plasma samples were collected before and at 20 min after the galactose bolus. The plasma was separated, deproteinized using perchloric acid, and analyzed for concentrations of galactose by the formation of NADH from NAD by galactose dehydrogenase (Boehringer Mannheim, Darmstadt, Germany). Following the galactose test, a 24-h enteral nutrition balance study was conducted for the two resected groups, where the pigs were given 7.5 ml/kg per h of bovine colostrum (590 kJ/kg body wt per day). Stoma output was collected throughout the balance period and quantified. After mixing all collections from each pig, we analyzed the energy content by bomb calorimetry and carbohydrate concentration by the Englyst method, the protein concentration by the Kjeldahl method, and the fat concentration by Van der Kamer titration, as previously described (22), with relative absorption calculated as: (intake − output)/(intake × 100). All animal procedures were approved by the Danish Animal Experimentation Inspectorate in accordance with the Council of Europe Convention ETS 123.

Tissue collection.

By the end of the 24-h enteral balance study, tissue protein synthesis was measured as described previously using stable isotopes (4). Briefly, piglets were given an intra-arterial infusion of l-phenylalanine (1.5 mmol/kg containing 0.15 mmol/kg l-[ring-¹³C₆] phenylalanine; Cambridge Isotopes Laboratories, Woburn, MA). Thirty minutes after isotope infusion, pigs were anesthetized, blood was drawn by cardiac puncture, and the pigs were euthanized with pentobarbitone (43). The small intestine, from the pyloric sphincter to the jejunostomy, was rapidly excised and placed on an ice-cold metal plate in a relaxed state. The total length was measured, and the small intestine was weighed. The remaining distal small intestine was sampled for histology and stereology using systematic uniform random sampling. A total of nine sections were taken throughout the intestine starting from a predetermined random start site and with a set sampling distance (total length/9). At each site, a 0.5-mm transverse biopsy was taken and fixed in 4% neutral buffered paraformaldehyde. Just proximal to the resection site, 10 cm of intestine was taken and placed in ice-cold Ringer for ex vivo nutrient absorption (see details below), and pieces from the proximal and distal segment were snap frozen in liquid nitrogen. In addition, 10 cm of duodenum and 10 cm of jejunum was removed and opened along the length, the mucosa was scraped off with a plastic slide, and the proportion of mucosa was determined after drying both the mucosa and the remaining intestine for 72 h at 60°C. The colon, heart, liver, lung, spleen, and stomach were weighed.

Intestinal protein synthesis.

Frozen tissue samples were homogenized and deproteinized using 2M perchloric acid, and the perchloric acid-soluble and the acid-insoluble fractions were analyzed by mass spectrometry. The acid-insoluble fraction was hydrolyzed using 6 M HCl for 24 h before gas chromatography-mass spectrometry analysis. In the two tissue pools, the isotopic enrichment of l-[ring-¹³C₆] phenylalanine] was determined by the N-propyl ester heptafluorobutyramide derivative using methane-negative chemical ionization. The analyses were performed on a 5890 series II gas chromatograph linked to a 5989B quadrupole mass spectrometer (Hewlett-Packard, Palo Alto, CA). Tissue samples were analyzed for protein content using the bicinchoninic acid method (Pierce, Rockford, IL).

Ex vivo absorption and mucosal enzyme activity.

One snap-frozen sample from the proximal and distal remnant small intestine from each pig was homogenized in 1.0% Triton X-100 (6 ml/g). The homogenates were assayed for disaccharidases (lactase, maltase, and sucrase) and aminopeptidase N, (ApN), aminopeptidase A (ApA), and DPP-4 activities, as previously described (36). The amount of substrate hydrolyzed in 1 min at 37°C was set to represent one unit of enzyme activity and was expressed per gram of intestine.

For ex vivo assessment of nutrient absorption, intestinal segments were obtained from the distal remnant small intestine at euthanasia. The tissue sections were kept in ice-cold Ringer solution before they were everted, cut into 1-cm pieces, and mounted on stainless steel rods. The sections were incubated in aerated Ringer nutrient solution for 4 min at 37°C. The capacity for nutrient uptake was measured for glucose and leucine, both at 5 and 50 mmol/l. The accumulation of nutrients by the mucosa was quantified by adding traceable amounts of either [U-14C] d-glucose (1.0 μl/10 ml) or l-[4,5–3H(N)]-leucine (2.0 μl/10 ml). The use of tracer-labeled d-glucose provides a sensitive indicator of the activities of the apical membrane glucose transporter, sodium-dependent glucose transporter 1. For leucine, the uptake represents both passive influx and active transport by carrier systems.

Plasma GLP-2 and amino acid measurements.

Plasma samples for GLP-2 measurements were taken at 0, 48, 96, and 120 h after surgery. All samples were extracted in a final concentration of 75% ethanol before GLP-2 measurements. GLP-2 was measured with an NH₂-terminal specific antiserum (code no. 92160), measuring only GLP-2 with an intact NH₂ terminus, as previously described (18). For standards, we used recombinant human GLP-2, and the tracer was rat GLP-2 with an Asp³³ -> Tyr³³ substitution. For amino acid concentration measurements, plasma samples were mixed with an equal volume of methionine sulfone (4 mM) as internal standard and centrifuged at 10,000 g for 30–60 min at 4°C through a 10-kDa cutoff filter. The filtrate was dried, and the amino acids were analyzed by reverse-phase high-pressure liquid chromatography of their phenylisothiocyanate derivatives (PicoTag; Waters, Woburn, MA) (42).

Histology, stereology, and Western blot.

The fixed samples were embedded in paraffin, cut into 5-μm thin paraffin sections, and stained with hematoxylin and eosin for morphometry measurements. An Olympus microscope running new CAST software (Visiopharm, Hørsholm, Denmark) was used for quantitative stereological analysis. The volume of the intestinal layers was estimated using the Cavalieri principle, volume = t * Ap * ∑p, where t is the distance between the sections, p is the number of points that fall on the region of interest, and A is the area associated with each test point in the point grid. The grid was designed to provide a total of 100–200 hits per region, which is sufficient to give a coefficient of error <5% (16). Villus height and crypt depth were determined on digital images acquired with an Axiophot microscope (Carl Zeiss, Oberkochen, Germany) and analyzed with NIH image software version 1.60 (SoftWoRx Explorer version 1.1; Applied Precision, Issaquah, WA). A total of 10–15 well-oriented villus and crypts, throughout the remnant intestine, were measured, and the values were averaged for each sample.

Slides were deparaffinized, endogenous binding sites were blocked in 0.6% H₂O₂, and antigen retrieval was conducted using heat-induced epitope retrieval in Tris-EGTA buffer (pH 8.0). Mouse monoclonal anti-KI67 (MIB-1; Dako, Glostrup, Denmark) was added (dilution 1:200) and incubated overnight at 4°C. The slides were developed after incubation in primary antibody enhancer, and addition of horseradish peroxidase polymer and the 3-amino-9-ethylcarbazole power vision were used for detection. Thereafter, the slides were counterstained in hematoxylin. The slides were counted using a stereological counting frame in newCAST (Visiopharm, Hørsholm, Denmark).

Equal amounts of protein samples were separated on SDS-PAGE, transferred to a nitrocellulose membrane, and incubated with primary antibody: cleaved caspase 3 (CC3), proliferating cell nuclear antigen (PCNA), and GAPDH overnight. This was followed by a 1-h incubation with secondary antibody, and the membranes were washed and developed with enhanced chemiluminescence (ECL-plus; Amersham, Piscataway, NJ), visualized using a ChemiDoc (UVP, Upland, CA).

Statistical analysis.

Time series data were analyzed by the MIXED procedure in SAS using a Gaussian spatial correlation (version 9.2; SAS institute, Cary NC). All other analyses were done using the statistical programming language R. Litter and pig were set as random variables, and treatment and section were set as explanatory variables. Post hoc comparisons were done with Wald testing with single-step P value adjustment for multiple comparisons. When residual variance heterogeneity was detected, the data were log-transformed before analysis. Data are reported as means ± SE with a P value <0.05 considered significant for all analyses.

RESULTS

Body weight and organ growth.

Relative to controls, the two resected groups showed a decreased weight gain from 3 days after resection (Fig. 1A and Table 1). Both SBS groups had a transient hypersecretion after surgery, but this resolved within the first days. Only transient episodes of hypoglycemia and alkalosis were seen during the experimental protocol. This was corrected by increasing TPN and sodium chloride to achieve normal plasma glucose (2.2 ± 0.3 mmol/l), pH (7.36 ± 0.03) and base excess (−1.2 ± 1.9 mmol/l) across groups.

Fig. 1.

A: development in body weight (BW) after resection. SBS, short bowel syndrome; TPN, total parenteral nutrition; EN, enteral nutrition. B and C: intestinal length and mass of remnant small intestine. D: plasma glucagon-like peptide (GLP)-2 concentrations before and after resection. E: in vivo enteral balance. F: ex vivo nutrient uptake of glucose and leucine. Data are presented as mean ± SE. Significant difference between groups (*P < 0.05; **P < 0.01; ***P < 0.001).

Table 1.

Body and organ weights

	Control	SBS	SBS+GLP-2
Birth body weight, g	993 ± 119	1130 ± 66	1100 ± 52
Postsurgical body weight, g	1017 ± 110	1068 ± 79	1044 ± 47
Final body weight, g	1385 ± 182	1207 ± 75	1103 ± 42
Average daily weight gain, g/kg per day	75 ± 16†	19 ± 6*	12 ± 9*
Resected intestinal length, cm/kg body wt	NA	137.6 ± 5.9	138.2 ± 4.2
Resected intestinal weight, g/kg body wt	NA	14.19 ± 0.48	14.15 ± 0.59
Proximal small intestinal mucosa, %	70.1 ± 2.4†	62.6 ± 2.7*	70.6 ± 1.6†
Colon, g/kg body wt	8.9 ± 1.9†	4.9 ± 0.7*	5.2 ± 0.8*
Stomach, g/kg body wt	5.0 ± 0.3	5.8 ± 0.6	6.5 ± 1.2
Liver, g/kg body wt	23.2 ± 0.6	26.6 ± 1.9	27.9 ± 2.4
Spleen, g/kg body wt	2.2 ± 0.3*	3.8 ± 0.7†	2.6 ± 0.3*†
Kidneys, g/kg body wt	8.7 ± 0.5*	9.8 ± 0.5*†	10.9 ± 0.51†
Lungs, g/kg body wt	24.2 ± 2.7	28.0 ± 2.7	26.3 ± 2.2
Heart, g/kg body wt	7.2 ± 0.6	8.1 ± 0.5	7.2 ± 0.8

Values mean ± SE (n = 5–8).

^*†Values significantly different (P < 0.05). SBS, short bowel syndrome; GLP, glucagon-like peptide; NA, not assessed.

At euthanasia, the relatively small intestinal weight and percentage of mucosa were increased in the SBS+GLP-2 group, compared with SBS (Fig. 1B, Table 1), with no significant change in relative length (Fig. 1C). The resected segment in the SBS group was similar in length to the segment excised at euthanasia, and the two resected groups had higher relative spleen weight and lower colon weights, compared with the control (Table 1).

Plasma GLP-2.

Before resection, the three groups had similar levels of circulating GLP-2 (∼45 pM, Fig. 1B). Within the first 2 days after resection and transition to TPN, GLP-2 values in the SBS group decreased (P < 0.001), and, from 48 h after resection and throughout the experiment, the GLP-2 values remained lower than controls (P < 0.05). During the enteral balance study, a significant increase in GLP-2 occurred in the control group but not in the two resected groups. The SBS+GLP-2 group had markedly elevated GLP-2 levels (P < 0.001) throughout the study (2,000–3,000 pM) (Fig. 1D).

In vivo nutrient absorption.

GLP-2 decreased the absolute stoma loss of fluids (wet weight), energy, and the individual macronutrients (P < 0.05, Table 2), leading to a higher relative absorption of fluid, energy, and all macronutrients in the SBS+GLP-2 group, relative to SBS (Fig. 1E). There was no significant difference in the absolute nutrient balances for either fluid or macronutrients. No differences were observed in the acute galactose absorption test with similar circulating levels of galactose 20 min after a galactose bolus (∼1.3 ± 0.2 mmol/l) in all groups. Ex vivo tissue nutrient uptake showed no differences in uptake of either glucose or leucine among the three treatment groups (Fig. 1F).

Table 2.

Twenty-four-hour enteral nutrient balance study

	SBS	SBS+GLP-2	P Value
Wet Weight, ml/day
Intake	197.2 ± 14.4	181.2 ± 6.5	0.27
Stoma output	153.2 ± 21.7	99.3 ± 15.8	0.04
Absolute absorption	44.2 ± 18.9	81.9 ± 13.0	0.09
Energy, kJ/day
Intake	788.0 ± 57.1	724.0 ± 25.3	0.28
Stoma output	279.4 ± 42.9	157.5 ± 24.6	0.01
Absolute absorption	508.6 ± 64.2	566.6 ± 21.8	0.36
Protein, kJ/day
Intake	376.1 ± 27.3	345.1 ± 12.3	0.27
Stoma output	151.5 ± 18.2	91.1 ± 11.4	0.01
Absolute absorption	224.6 ± 26.7	254.0 ± 11.4	0.29
Fat, kJ/day
Intake	299.5 ± 21.7	274.8 ± 9.8	0.27
Stoma output	105.7 ± 21.7	51.8 ± 10.9	0.02
Absolute absorption	193.7 ± 29.0	222.0 ± 10.4	0.31
Carbohydrate, kJ/day
Intake	93.6 ± 7.0	85.9 ± 3.3	0.29
Stoma output	13.1 ± 2.1	8.5 ± 0.9	0.03
Absolute absorption	80.6 ± 6.3	77.4 ± 3.0	0.63

Values are means ± SE, n = 6–7.

Intestinal protein synthesis and plasma amino acids.

By the end of the experiment, the GLP-2 treatment had increased the small intestinal DNA and protein content to levels similar to those in intact controls, and both GLP-2+SBS and controls had a high level of intestinal DNA (160 vs. 86 mg/kg body wt, P < 0.05) and protein content (3.2–3.5 vs. 1.9 g/kg body wt, P < 0.05). Both resected groups had lower FSR than unresected control pigs (Fig. 2A), but, calculated at absolute rates of protein synthesis, GLP-2 increased protein synthesis values, relative to SBS alone (Fig. 2D). Increased stomach protein content was also observed in the SBS+GLP-2 compared with the other groups (Table 3).

Fig. 2.

A and D: fractional and absolute protein synthesis rate in remnant intestine. B, C, E, and F: plasma levels of citrulline, ornithine, arginine, and essential amino acids, respectively. Data are presented as means ± SE. Significant difference between groups (*P < 0.05; **P < 0.01; ***P < 0.001).

Table 3.

DNA, protein concentration, FSR, and ASR in the stomach, liver, and muscle

	Control	SBS	SBS+GLP-2
Stomach
DNA content, mg/kg body wt	29.0 ± 3.9	27.4 ± 4.1	34.8 ± 2.6
Protein content, g/kg body wt	0.45 ± 0.0*	0.52 ± 0.1*	0.69 ± 0.1†
FSR, %/day	34.0 ± 1.9	32.5 ± 3.5	34.4 ± 6.8
ASR, g/kg body wt per day	0.14 ± 0.0	0.17 ± 0.0	0.24 ± 0.1
Liver
DNA content, mg/kg body wt	110.5 ± 4.0	126.7 ± 4.9	119.8 ± 8.3
Protein, g/kg body wt	3.3 ± 0.1	4.00 ± 0.3	4.3 ± 0.4
FSR, %/day	53.9 ± 3.9	59.8 ± 5.6	58.5 ± 3.0
ASR, g/kg body wt per day	1.8 ± 0.1	2.5 ± 0.5	2.5 ± 0.4
Muscle
DNA concentration, μg/mg	4.1 ± 0.3	3.9 ± 0.2	4.4 ± 0.2
Protein concentration, μg/mg	84.2 ± 2.7	88.9 ± 4.7	91.9 ± 3.4
FSR, %/day	18.0 ± 4.2	17.5 ± 6.8	17.9 ± 3.6

Values mean ± SE (n = 5–8).

^*†Values are significantly different (P < 0.05). FSR, fractional synthesis rate; ASR, absolute synthesis rate.

GLP-2 induced a marked upregulation of circulating levels of citrulline, arginine, and ornithine to the levels of unoperated pigs (Fig. 2, B, C, and E). This upregulation was specific for the citrulline-arginine-ornithine cycle, and the two SBS groups showed marked decreases in essential amino acids compared with controls (Fig. 2F), with no effect of GLP-2. A similar pattern was observed for nonessential amino acids (Table 4).

Table 4.

Free amino acids in plasma (μM)

	Control	SBS	SBS+GLP-2
Alanine	414.7 ± 44.0	433.9 ± 56.9	527.7 ± 162.8
Asparagine	179.4 ± 39.7†	48.7 ± 12.3*	73.0 ± 8.4*
Aspartic acid	56.0 ± 9.7	39.7 ± 8.7	36.2 ± 4.1
Cysteine	902.7 ± 101.0	943.4 ± 94.8	1106.4 ± 232.8
Cystine	22.2 ± 6.9	21.5 ± 4.1	21.4 ± 4.9
Glutamine	269.7 ± 30.1	194.3 ± 25.2	225.8 ± 79.5
Glutamic acid	143.3 ± 12.3	163.59 ± 14.7	126.1 ± 13.6
Glycine	913.8 ± 58.7	611.8 ± 113.1	673.6 ± 136.3
Histidine	72.0 ± 8.8	50.6 ± 3.3	54.4 ± 13.2
Isoleucine	244.0 ± 34.9†	143.1 ± 10.6*	159.5 ± 13.2*
Leucine	341.7 ± 46.0†	205.6 ± 12.9*	232.4 ± 24.73*
Lysine	328.4 ± 47.5†	179.7 ± 17.3*	232.8 ± 53.2*†
Methionine	143.7 ± 19.4†	70.9 ± 13.7*	83.4 ± 14.3*†
Norleucine	386.1 ± 7.3	396.4 ± 4.6	382.5 ± 6.8
Proline	729.9 ± 84.5†	324.3 ± 47.7*	481.9 ± 116.9*†
Serine	402.6 ± 36.9†	170.8 ± 23.9*	222.8 ± 33.5*
Taurine	148.5 ± 32.1	141.6 ± 13.4	276.6 ± 72.9
Threonine	600.9 ± 123.8†	240.5 ± 50.8*	266.2 ± 49.9*
Tryptophan	33.8 ± 3.9†	23.8 ± 2.6*	20.7 ± 3.5*
Tyrosine	366.9 ± 24.4†	224.3 ± 39.8*	213.6 ± 42.3*
Valine	726.8 ± 84.4†	394.2 ± 24.4*	467.6 ± 37.4*

Values are means ± SE.

^*†Values are significantly different.

Ex vivo digestive enzymes and nutrient absorption.

We next examined functional parameters, and intestinal resection did not alter the activity of brush border enzymes, relative to unresected control pigs (Fig. 3). However, GLP-2 induced a marked increase in the activity of brush border enzymes, especially in the proximal part of the remnant intestine (Fig. 3) relative to SBS. This was most pronounced for sucrase and maltase activities (+300%, Fig. 3, A and B). GLP-2 induced a marginal increase in proximal lactase activity (Fig. 3C). GLP-2 induced ApN and DPP-4 activities (+200%, Fig. 3, C and D), whereas there was no effect on ApA activity (Fig. 3F).

Fig. 3.

Activity of digestive enzymes in the proximal and distal remnant small intestine. A–C: activity of disaccharidases (sucrase, maltase, lactase). D–F: activity of peptidases [dipeptidyl peptidase-4 (DPPIV), aminopeptidase N (ApN) and aminopeptidase A (ApA)]. Data are presented as means ± SE. Significant difference between groups (*P < 0.05; ***P < 0.001).

Histology and Western blot.

In parallel with the changes in gut mass, we observed an increase in total small intestinal volume in SBS-GLP-2-treated neonates compared with SBS (Fig. 4A). Subanalyses of the cellular layers revealed further that this difference was mainly driven by an increased mucosa volume, as no significant changes were observed in the submucosa, serosa, or muscularis layers (Fig. 4A). The marked increase in mucosa volume is probably related to an increased in villus surface area as indicated by the observed differences in villus height and crypt depth in the SBS+GLP-2 compared with the other groups (Fig. 4, B and C). No significant differences in total number of proliferative cells were seen among groups (data not shown), and a small difference was seen in number of Ki67-positive cells per crypt area between controls and SBS+GLP-2 (Fig. 4D). These findings were supported by Western blot for CC3 and PCNA, which showed no difference (Fig. 4E).

Fig. 4.

A–C: small intestinal volume, villus height, and crypt depth. D: Ki67-positive cells per crypt area. E: Western blot of proliferating cell nuclear antigen (PCNA) and cleaved caspase 3. F: representative hematoxylin-eosin sections of the small intestine in control, SBS, and SBS+GLP-2, respectively. Significant difference between groups (*P < 0.05; ***P < 0.001).

DISCUSSION

SBS is a well-known complication following intestinal resection in infants suffering from congenital anomalies or NEC (47). In adult patients with SBS, exogenous GLP-2 improves intestinal adaptation and stimulates intestinal function (20), but evidence is lacking regarding the effects of GLP-2 in SBS newborns. GLP-2 stimulates growth and function of the intact intestine of TPN-fed preterm and term pigs (30, 31), but the efficacy appears diminished in preterm pigs (34). In this study, we combined an immature state of the intestine with SBS and jejunostomy in preterm pigs. We document that supply of exogenous GLP-2 in the days just after resection induces marked improvements in intestinal adaptation and function.

A commonly used clinical index of intestinal function in patients with SBS is the net enteral nutrient uptake, which can diminish the need for parenteral nutrition. We found that GLP-2 treatment decreased the daily stoma losses of fluid and macronutrients, which translated into significantly increased relative nutrient absorption of fluid and all macronutrients. In principle, the GLP-2-induced improvement in relative absorption could result from direct antisecretory or antidiarrheal effects. The findings that GLP-2 increases the relative absorption of especially fluid and protein are also supported by clinical studies in adult patients with SBS (20), whereas the effects on enteral balance of antisecretory agents (e.g., ranitidine) are primarily related to fluid (25). Our ex vivo data for nutrient absorption and digestive enzymes suggest that the improved nutrient balance is likely due to increased digestive capacity, rather than transport-mediated nutrient uptake per se. Our data suggest that the important effect seen in enteral nutritional balance is mediated via the combined effect of a marked increase in functional tissue mass and a stimulation of some functions of intestinal tissue, such as (e.g., brush border enzyme activities) but not hexose absorption capacity. Collectively, the results in this study on preterm SBS piglets suggest that GLP-2 can be used to improve nutrient availability, fluid loss, and possibly growth in the highly vulnerable infant patients with SBS group. Continuous infusion of the native peptide with TPN was sufficient to mediate the effect in preterm pigs, but it remains to be tested whether long-acting GLP-2 analogs or GLP-2 combined with early minimal enteral nutrition will have the potential to exhibit the same effects in this model and possibly in infants too.

In our model, a jejunostomy was made to specifically study nutrient balance without a functioning colon, which reflects the clinical situation of newly resected infants with limited colon stimulation and thus endogenous GLP-2 secretion. The insufficient intestinal adaptation in preterm SBS pigs likely reflects insufficient endogenous GLP-2 secretion with much lower values in the SBS pigs (14 ± 2 pM), relative to the unresected control pigs (34 ± 3 pM) or intact TPN-fed term piglets (35 ± 4 pM) observed previously (3). After introduction of enteral nutrition during the balance study period, the SBS pigs showed no increase in GLP-2 (19 ± 3 pM), whereas the levels increased sharply in the control pigs (63 ± 9 pM, Fig. 1B), indicating that SBS preterm pigs were unable to mount a GLP-2 secretion response to enteral feeding. In more mature rodents (27), pigs (29), and humans (24), jejunal resection with a reconnected colon leads to elevated plasma GLP-2 levels. Even if the ileum is resected and the colon remains in continuity with the remnant intestine, GLP-2 levels are still increased (21), and this endogenous colonic GLP-2 release may significantly improve intestinal adaptation (6, 26, 44). Our results suggest that, under the clinically relevant conditions of SBS with a jejunostomy, endogenous GLP-2 deficiency leads to limited adaptation of the remnant intestine in the premature neonatal intestine. In addition, the limited adaptation may be explained by the immature state of the preterm intestine. We have previously documented enhanced intestinal adaptation in term SBS pigs with a jejunostomy, relative to preterm SBS pigs (38).

Our results show that exogenous GLP-2 partly restored intestinal structure and function in preterm pigs with SBS and a jejunostomy. The pharmacological increase in circulating GLP-2 levels resulted in a remarkable increase in the total intestinal mass and protein and DNA content of the remnant intestine to levels that approached the value in unoperated pigs. The changes in tissue mass were matched by increases in volume of mucosa, villus height, and crypt depth, indicating rapid structural changes associated with GLP-2 treatment similar to or even more pronounced than those reported in adult SBS models (10, 27). The changes were supported by parallel increases in protein synthesis, like in intact newborn pigs (3, 4), and increased circulating arginine, ornithine, and citrulline levels. The increase in epithelial cell mass leads to increased plasma citrulline, which is further metabolized to arginine (13, 45) despite that total levels of essential amino acids were decreased in both resection groups. This may be explained by the fact that protein absorption was only 60–80% of oral supply, similar to observations from infants with a jejunostomy (39). This may lead to lower circulating free amino acids, even with GLP-2 stimulation of absorption.

Despite the increased intestinal mucosal villus height, crypt depth, and total volume, we did not see an increased proliferation after GLP-2 treatment, contrasting some other reports (3, 4, 27, 29). However, the results are in agreement with previous findings in preterm TPN-fed pigs, where increased villus height was not associated with increased proliferation (30). Ki67 staining is a static marker of crypt cell proliferation, and exogenous GLP-2 may have stimulated cell proliferation and fractional rates of total protein synthesis mainly during the first days following resection; secondly high basal proliferation following enteral nutrition in preterms may hide GLP-2 effects. A detailed description of the specific SBS-and GLP-2-mediated changes in intestinal proteins will be presented as a separate report based on the gel-based proteomics analyses of intestinal tissues from the present study.

The trophic response to GLP-2 translated into significant improvements in clinically relevant indices of intestinal absorptive function. The increase in brush border enzyme activities, particularly sucrase and maltase in GLP-2-treated pigs, is in agreement with previous findings in rodents (40) and unresected preterm and term pigs (30, 31). On the other hand, GLP-2 had no effect on tissue-specific nutrient (glucose, leucine) transport and in vivo galactose uptake, which is consistent with our previous findings, showing hyporesponsiveness in fetal and preterm compared with term piglets (34, 37, 38).

Further studies are needed to investigate whether it would be possible to utilize the increased relative enteral absorption, reported in the current study, to decrease the parenteral support, as has been shown with GLP-2 to adults (19). The ultimate goal would of course be to wean neonates with intestinal resection off TPN, but more long-term studies are needed to investigate this. Furthermore studies are warranted to investigate whether various GLP-2 treatment regimens, with or without concurrent enteral nutrition, can augment intestinal adaptation and function to treat intestinal failure in pediatric patients.

In conclusion, we show novel evidence in parenterally fed, preterm piglets in which SBS with a jejunostomy induces a GLP-2-deficient state and limited intestinal adaptation. We also show that GLP-2 treatment effectively induces intestinal adaptation, even under conditions of exclusive parenteral nutrition. Moreover, the intestinal trophic effect of 5 days of GLP-2 treatment translated into an increased intestinal absorptive function during the first 24 h of enteral feeding after resection. These results extend our previous evidence that GLP-2 exerts trophic effects in preterm neonates and suggest that GLP-2 treatment is a promising therapy to maintain intestinal function in neonatal patients immediately after intestinal resection.

GRANTS

This study was supported by a grant from the Danish Strategic Research Council as well as the UNIK programme.

DISCLOSURES

Dr. Palle B. Jeppesen is a paid consultant for NPS Pharmaceuticals and Nycomed, now Takeda Pharmaceuticals. All other authors have no conflict of interest to declare.

AUTHOR CONTRIBUTIONS

Author contributions: A.V., T.T., J.J., D.G.B., P.B.J., and P.T.S. conception and design of research; A.V., T.T., P.L., B.S., S.B.B., B.H., N.Q., and P.T.S. performed experiments; A.V., P.L., B.S., S.B.B., B.H., J.J., and D.G.B. analyzed data; A.V., T.T., P.L., B.S., S.B.B., B.H., J.J., N.Q., D.G.B., P.B.J., J.J.H., and P.T.S. interpreted results of experiments; A.V. prepared Figs.; A.V. drafted manuscript; A.V., T.T., P.L., S.B.B., B.H., J.J., N.Q., D.G.B., P.B.J., J.J.H., and P.T.S. edited and revised manuscript; A.V., T.T., P.L., B.S., S.B.B., B.H., J.J., N.Q., D.G.B., P.B.J., J.J.H., and P.T.S. approved final version of manuscript.