IGF Binding Protein-4 is Required for the Growth Effects of Glucagon-Like Peptide-2 in Murine Intestine

Abstract

Glucagon-like peptide-2 (GLP-2) is an enteroendocrine hormone that stimulates the growth of the intestinal epithelium. We have previously demonstrated that GLP-2 exerts its intestinotropic effect through an indirect mechanism that requires both IGF-1 and the intestinal epithelial IGF-1 receptor. However, the biological activity of IGF-1 is modulated by IGF binding proteins (IGFBPs), including IGFBP-4, which is highly expressed in the intestine. To determine the role of IGFBP-4 in the tropic effects of GLP-2, IGFBP-4 knockout (KO) and control mice were treated with degradation-resistant GLP-2 or vehicle for 10 days. Comparable levels of IGFBP-1–3/5–7 mRNAs were observed in the intestinal mucosa of all animals. IGFBP-4 KO mice had greater small intestinal weight and length, and deeper crypts (P < .05) as compared with controls, suggesting that IGFBP-4 has an inhibitory role in basal intestinal growth. However, small intestinal weight, crypt-villus height and crypt cell proliferation increased in response to GLP-2 in control mice (P < .05), and these changes were abrogated with IGFBP-4 KO. In contrast, pregnancy-associated plasma protein-A KO mice, which have increased levels of circulating IGFBP-4, demonstrated a normal intestinotropic response to GLP-2. Finally, GLP-2 treatment of control mice significantly increased IGFBP-4 mRNA expression in the jejunal mucosa (P < .05), a finding that was recapitulated by GLP-2 treatment of fetal rat intestinal cells in culture (10⁻⁸M for 2 h; P < .05). Collectively, these results indicate that the IGF-I-modulating protein, IGFBP-4, exerts a negative effect on basal intestinal growth but plays a positive regulatory role in the intestinotropic actions of GLP-2.

Glucagon-like peptide-2 (GLP-2) is a 33-amino acid endocrine hormone that is released from the intestinal L-cell in response to nutrient ingestion (1). GLP-2 has many beneficial effects in the gut, including stimulation of intestinal epithelial cell proliferation (1, 2), as well as enhancing the barrier function of the epithelium and improving both nutrient absorption and blood flow (1–5). The long-acting analog of GLP-2, teduglutide, has recently been approved for use in patients with short bowel syndrome (6). However, the mechanisms underlying the tropic actions of GLP-2 on the gut are complex and not well defined. The epithelial layer of the intestine, where the effects of GLP-2 are observed, does not express the GLP-2 receptor (GLP-2R); rather, previous studies have demonstrated that GLP-2R expression is localized to the subepithelial layer of the mucosa (7, 8), as well as rare enteroendocrine cells (9) and scattered enteric neurons (10). Thus, it is believed that GLP-2 exerts its beneficial effects on the epithelium through secondary mediators. For example, keratinocyte growth factor (8) and epidermal growth factor (11) have both been reported to be involved in the intestinotropic effects of GLP-2. However, our laboratory has established that the growth and barrier effects of GLP-2 are mediated through a pathway that requires IGF-1, but not IGF-II, as well as the type 1 IGF receptor expressed specifically on the intestinal epithelial cells (Intestinal epithelial IGF-I receptor; IE-IGF-IR) (3, 12, 13).

Circulating IGF-1 is a whole-body growth factor that also induces epithelial proliferation and intestinal growth and promotes growth in intestinal resection and parenteral feeding models (14, 15).The bioactivity of IGF-1 is known to be regulated by IGF binding proteins (IGFBPs), which act as reservoirs for circulating IGF-I, thereby prolonging its half-life (16). Conversely, IGFBPs may also interfere with IGF-1 signaling, thereby inhibiting the downstream effects of this growth factor (17). There are 6 main IGFBP isoforms within the IGFBP superfamily, as well as a number of other related proteins (18, 19). Within the intestine, IGFBP-3, IGFBP-4, and IGFBP-5 are most highly expressed (20). Although a previous study has shown that IGFBP-3/5 double knockout (KO) mice retain their proliferative response to GLP-2 (21), the role of IGFBP-4 in GLP-2-induced intestinal growth has not been studied to date. Furthermore, both positive and negative roles of IGFBP-4 on IGF signaling have been reported for other tissues, such that overexpression of IGFBP-4 inhibits IGF action on the neointima (22), whereas KO of IGFBP-4 conversely impairs fetal growth (23).

The mechanisms by which IGFBPs regulate the bioactivity of IGF-1 in vivo are made further complex by the existence of IGFBP-cleaving enzymes. Although IGFBP-4 can be degraded by nonspecific enzymes (16), IGFBP-4-specific proteases, such as pregnancy-associated plasma protein-A (PAPP-A), have also been described (24). PAPP-A protease activity is IGF dependent; in other words, binding of IGF-1 (or IGF-II) to IGFBP-4 greatly increases the rate of IGFBP-4 cleavage by PAPP-A (25). Interestingly, in some tissues, PAPP-A potentiates IGF signaling in a paracrine manner, by cleavage of IGFBP-4, thereby liberating bound IGF (26). Conversely, KO of PAPP-A in mice results in increased circulating levels of IGFBP-4 in association with a decrease in whole-body growth (23). However, to date, no role for PAPP-A in intestinal physiology has been described.

To determine whether IGFBP-4 plays a role in the IGF-I-IE-IGF-IR axis that mediates the intestinal growth effects of GLP-2, we used IGFBP-4 null mice as a whole-body loss-of-function model. PAPP-A KO mice were also used as a gain-of-function model, to control for the effects of circulating IGFBP-4 on the intestine. The results of these studies indicate that the absence of IGFBP-4 impairs the epithelial proliferative and intestinal growth responses to GLP-2 and that these actions of IGFBP-4 are likely to be mediated in a paracrine manner.

Materials and Methods

Animals

Animals were kept in an animal facility at the University of Toronto with ad libitum feeding under a 12-hour light, 12-hour dark cycle. All animal protocols were approved by the Animal Care Committee of the University of Toronto. The IGFBP-4 KO mice have previously been reported (23). Control mice were a mixture of wild-type (WT) littermates and C57BL/6 mice (Charles River Canada); no experimental differences were observed between these different groups of animals (data not shown). PAPP-A KO mice and WT littermate controls on a C57BL/6, 129 background (27) were a kind gift from Dr Cheryl A. Conover (Mayo Clinic, Rochester, MN). All mice used for in vivo studies were age matched (8–10 wk) and included both sexes. Genotypes were confirmed by PCR amplification of DNA extracted from ear clips using primers in Supplemental Table 1. In vitro studies were conducted using fetal rat intestines collected from pregnant female Wistar rats at gestational day 19–20 (Charles River).

In vivo studies

The degradation-resistant GLP-2 analog, h(Gly²)-GLP-2 (GLP-2) (3, 11–13, 28), was purchased from American Peptide Company, Inc. As described previously (3, 12, 13), mice were given sc injections of either vehicle (PBS), pharmacological (0.1 μg/g) or suprapharmacological dose (10 μg/g) of GLP-2, daily for 10 days with a final injection 3 hours before euthanasia. The mice were weighed, intestines were collected, and whole intestinal lengths were measured vertically under constant tension. The intestines were then flushed with PBS, blotted, and weighed. Two 2-cm segments of jejunum were collected from each sample, flash-frozen, and stored at −80°C for RNA extraction from mucosal scrapes or whole thickness. One 2-cm segment of jejunum was collected from each sample, fixed in 10% formalin, and paraffin embedded for immunohistochemical and morphometric analyses.

In vitro studies

As described and validated previously (12, 29), fetal rat intestinal cells (FRICs) were collected by enzymatical digestion of whole fetal rat intestines (mixed sex) and cultured overnight in DMEM containing 4.5-g/L glucose, 5% fetal bovine serum, and 40-U/mL penicillin and 40-μg/mL streptomycin in 10-cm plates. Cells were then treated with GLP-2 (10⁻⁸M) or vehicle in low-glucose DMEM with 0.5% FBS for 0.5–24 hours.

Gene expression

Total RNA from whole intestine, mucosal scrapes, or FRIC cultures was extracted using the RNeasy kit (QIAGEN, Inc). Reverse transcription was conducted using 5X All-in-One RT MasterMix (Applied Biological Materials, Inc), and semiquantitative PCR was conducted using TaqMan Gene Expression assay (Life Technologies, Inc) with primers found in Supplemental Table 2 and 18S as the internal control as previously validated (12). Expression of the target gene was calculated using the delta-delta cycle-threshold method.

Microscopy

For morphometric analysis and immunohistochemistry, 4 transverse cross-sections of each jejunal sample were used. For morphometric analysis, sections were stained with hematoxylin and eosin and imaged with a Zeiss microscope and analyzed using AxioVision Version 4.9.1 software (Carl Zeiss). Crypt depth was measured from the crypt base to the villus junction and villus height measured from tip of the villus to the villus junction. Crypt-to-villus height was calculated as the sum of the mean crypt depth and mean villus height for each mouse. At least 20 measurements were taken for each crypt and villus measurement per mouse for the IGFBP-4 mice. However, the PAPP-A mouse intestines were found to be very fragile upon section cutting, and a minimum of 10 crypts and villi were therefore measured for each mouse. Muscle thickness was determined at 8 individual points around each cross-sectional circumference for 2–4 cross-sections per mouse. All measurements were done in a blinded manner.

Immunohistochemistry for the proliferative marker Ki-67 (monoclonal antibody #Ki-67) was conducted using rabbit antihuman Ki-67 antibody (see Antibody Table) (Abcam). In brief, antigen retrieval was performed in citrate buffer and nonspecific binding was blocked using 5% normal goat serum. Endogenous peroxidase activity was blocked using 3% H₂O₂, and slides were incubated with the primary antibody overnight, followed by a goat antirabbit secondary antibody (see Antibody Table) (Vector Laboratories, Inc). Negative controls were performed without the primary antibody (data not shown). Visualization was conducted using Ultra Streptavidin-Horseradish Peroxidase Complex: Level 2 labeling reagent (Signet) and Sigma Fast-3,3′ Diaminobenzidine (Sigma-Aldrich, Inc). The percentage of Ki-67-positive cells was determined by recording Ki-67 positive and negative cells at each cell position, counting up 1 side of the crypt to cell position 22, for at least 20 crypts per mouse. All analyses were conducted in a blinded manner. Apoptosis was similarly assessed for the intestinal villi using the Apoptag Peroxidase In Situ Apoptosis Detection kit (EMD Millipore).

Statistical analyses

All data are presented as mean ± SEM. The results were analyzed using the Student's unpaired t test or two-way ANOVA followed by Student's t test. Some data were transformed using the Log₁₀ function to normalize variance. Significance was set at P < .05.

Results

To determine the role of IGFBP-4 in GLP-2-stimulated intestinal growth, mice were subjected to chronic treatment with either 0.01-μg/g GLP-2 or vehicle alone. Vehicle-treated IGFBP-4 KO mice were not significantly different from control animals with respect to body weight (Figure 1A) or small intestinal weight (Figure 1B). The intestinal weight of control mice treated with GLP-2 was increased by 28.0 ± 8.7% vs vehicle-treated animals (P < .05) (Figure 1B), as expected. However, this growth response to GLP-2 was impaired in IGFBP-4 KO mice (P > .05). In order to account for the sex differences in body weight and associated differences in intestinal size, small intestinal weight was also normalized to body weight (Figure 1C). Surprisingly, the small intestinal weight per body weight of vehicle-treated IGFBP-4 KO mice was found to be increased, by 30.6 ± 11.5% (P < .05), compared with control mice; this change was independent of normalized intestinal length, which was not different between the groups of animals (Figure 1D). Muscle thickness did not demonstrate any differences between mice (data not shown).

Figure 1.

IGFBP-4 KO mice have impaired intestinal growth responses to GLP-2. IGFBP-4 KO (KO) and control (CON) mice were treated with 0.1-μg/g hGly2-GLP-2 (GLP-2) or vehicle (PBS) for 10 days. Body weight (A), whole small intestinal weight (B), small intestinal length normalized to body weight (C), small intestinal weight normalized to body weight (D), and small intestinal weight normalized to whole small intestinal length (E); n = 10–12 per group. Open bar indicates vehicle treatment, and closed bar indicates GLP-2 treatment. *, P < .05 as indicated.

Consistent with the absolute small intestinal weight data, GLP-2-treated control mice demonstrated a 24.4 ± 8.6% (P < .05) and 23.4 ± 6.9% (P < .05) increase in small intestine weight compared with vehicle-treated control mice when normalized to body weight or intestinal length, respectively, whereas the growth response to GLP-2 was again found to be impaired in KO mice (P > .05) (Figure 1, C and E). This lack of a significant growth response to GLP-2 in KO mice was also found when the data were expressed as a fold of the paired PBS-treated animals (Supplemental Figure 1). To ensure that this lack of response to GLP-2 in the IGFBP-4 KO mice was not due to the small intestine having reached its maximal growth capacity, a preliminary chronic treatment study was conducted using a dose that was 100 times higher than the pharmacological dose initially tested (ie, 10 μg/g) (Supplemental Figure 2). For both the WT and KO animals, the intestinal weight increased overall by 60%–70%. This was greater than that seen using the lower dose of GLP-2 and indicates that the lack of response to the lower dose of GLP-2 in the IGFBP-4 KO mice was not due to any limitation in intestinal growth.

To determine whether the absence of IGFBP-4 impaired the known intestinal proliferative effect of GLP-2 (12, 13), jejunal sections of mice were stained with the proliferative marker, Ki-67, and subjected to positional analysis. Control mice displayed a significant increase in the percentage of Ki-67-positive cells at cell positions 12–14 (P < .05-.001) (Figure 2A) with GLP-2 treatment, whereas IGFBP- 4 KO mice demonstrated no change in the percentage of Ki-67-positive cells at any position within the jejunal crypt (P > .05). To compare between genotypes, the area under the curve (AUC) from cell position 12–17 was calculated for each group, because this corresponds to the region that is known to be sensitive to the proliferative effects of GLP-2 (12, 13). Consistent with the intestinal weight and proliferative data, the proliferative AUC increased by 96.3 ± 27.3% (P < .01) in GLP-2-treated control mice as compared with vehicle controls. However, despite an increase in the basal proliferative AUC in vehicle-treated IGFBP-4 KO mice (by 81.9 ± 35.8%; P < .05) as compared with vehicle-treated control mice, the KO mice displayed an abrogated proliferative AUC response to GLP-2 treatment (P > .05) (Figure 2C). Although slides were also stained for apoptotic cells, the number of cells per villus was too low to be reliably quantitated (data not shown). Finally, to determine whether the proliferative differences translated to morphometric changes in the intestinal epithelium, crypt depth and villus height were measured. As expected based on the gravimetric and proliferative responses, the crypt-to-villus height of control mice increased by 28.4 ± 7.4% (P < .05) with GLP-2 treatment, whereas this response was impaired in the IGFBP-4 KO animals (P > .05) (Figure 2D).

Figure 2.

IGFBP-4 KO mice have impaired intestinal proliferative responses to GLP-2. IGFBP-4 KO (KO) and control (CON) mice were treated with 0.1-μg/g hGly2-GLP-2 (GLP-2) or vehicle (PBS) for 10 days. A and B, Percentage of Ki-67-positive cells within the jejunal crypt for each cell position from the base to cell position 22 for (A) CON mice and (B) KO mice; n = 7–8 per group. C, AUC between cell positions 12 and 17 for CON and KO mice shown in A and B; n = 7–8 per group. D, Crypt depth and villus height measurements for CON and KO mice; n = 4–6 per group. Open bar indicates vehicle treatment, and closed bar indicates GLP-2 treatment. *, P < .05; **, P < .01; ***, P < .001 as indicated.

To ensure that the observed phenotype in the IGFBP-4 KO mice was not consequent to alterations in other intestinal IGFBP isoforms, the expression of IGFBP-1–3 and IGFBP-5–7 was determined in jejunal mucosal scrapes by semiquantitative PCR. IGFBP-1 and IGFBP-2 mRNA expression was not detectable in the mucosa, and levels were extremely low overall in whole jejunum; mucosal expression of the IGFBP-4 cleaving enzyme, PAPP-A, was also extremely low (data not shown). Furthermore, although mRNA for IGFBP-3 and IGFBP-5–7 was readily detected in whole jejunum, their relative expression in the mucosa was much lower (Figure 3). Notwithstanding, there were no significant differences in the expression of any of these IGFBP's between control and IGFBP-4 KO mice, with and without GLP-2 treatment.

Figure 3.

Intestinal IGFBP isoform IGFBP-3, IGFBP-5, IGFBP-6, or IGFBP-7 expression does not differ between mice. IGFBP-4 KO (KO) and control (CON) mice were treated with 0.1-μg/g hGly2-GLP-2 (GLP-2) or vehicle (PBS) for 10 days. mRNA expression of IGFBP isoform (A) IGFBP-3, (B) IGFBP-5, (C) IGFBP-6, and (D) IGFBP-7, as assessed by semiquantitative PCR of mucosal scrapes extracted from CON and KO treated with 0.1-μg/g GLP-2 or PBS for 10 days; n = 3–5 per group. Adjacent bar graph in each panel indicates relative mRNA expression in whole, jejunum of control mice; n = 4 per group. Open bar indicates vehicle treatment, and closed bar indicates GLP-2 treatment.

To investigate the role of circulating IGFBP-4 on GLP-2-induced intestinal growth, we conducted a chronic study using PAPP-A KO model as a gain-of-function model for circulating levels of IGFBP-4 (27). IGFBP-4 mRNA expression within the whole intestine of PAPP-A KO was comparable with that in WT mice (1.1 ± 0.3-fold of control; P > .05). However, as previously reported (23, 27), PAPP-A KO mice were smaller by 36.7 ± 3.9% (P < .01) compared with WT animals (Figure 4A). In addition to impaired overall growth, the small intestinal weight of PAPP-A KO mice was reduced by 32.4 ± 3.5% (P < .001) (Figure 4B). As a consequence, the small intestinal weight per body weight of vehicle-treated KO mice was not different from that of the vehicle-treated WT animals (Figure 1C). However, normalized intestinal length was unexpectedly increased by 42.4 ± 6.9% in the vehicle-treated KO mice (P < .05) (Figure 4D), resulting in a decrease in small intestinal weight per unit intestinal length, by 25.1 ± 2.9% (P < .001) (Figure 1E). As expected, GLP-2 stimulated an increase in intestinal growth of 20.8 ± 5.1% and 17.9 ± 3.2% in WT mice when intestinal weight was normalized to body weight and intestinal length, respectively, and similar increases of 21.6 ± 4.5% (P < .01) and 18.4 ± 3.9% (P < .01), respectively, were observed in the KO animals (Figure 4, C and E). Consistent with these findings, GLP-2 also stimulated an increases in the crypt-to-villus height, by 29.4 ± 5.4% and 11.5 ± 5.7% in WT and PAPP-A KO mice, respectively (P < .01 and P < .05) (Figure 4F).

Figure 4.

PAPP-A KO mice have preserved intestinal growth responses to GLP-2. PAPP-A KO (KO) and WT mice were treated with 0.1-μg/g hGly2-GLP-2 (GLP-2) or vehicle (PBS) for 10 days. Body weight (A), whole small intestinal weight (B), intestinal length normalized to body weight (C), intestinal weight normalized to body weight (D); n = 8–9 per group, and crypt depth and villus height (E); n = 5–8 per group. Open bar indicates vehicle treatment, and closed bar indicates GLP-2 treatment. *, P < .05; **, P < .01; or ***, P < .001 as indicated.

Finally, to establish whether jejunal IGFBP-4 mRNA expression is regulated locally by GLP-2, mucosal scrapes from control animals subjected to chronic treatment with GLP-2 were examined for changes in IGFBP-4 mRNA expression. GLP-2-treated animals demonstrated a 518 ± 22% increase (P < .05) in mucosal IGFBP-4 mRNA levels as compared with vehicle-treated controls (Figure 5A). Because the mucosal expression of IGFBP-4 transcripts was found to be very low relative to that of the whole intestine (Figure 5A), FRIC cultures were used as an in vitro model of the entire intestine. FRIC cultures have previously been established to express a functional GLP-2R that displays a cAMP response, as well as enhanced IGF-1 mRNA expression and IGF-1 secretion in response to GLP-2 treatment (12). When incubated with GLP-2 (10⁻⁸M) for 2 hours, IGFBP-4 mRNA expression in the FRIC cultures was also found to be increased, by 40.8 ± 15.2% (P < .05), compared with vehicle-treated cells (Figure 5B). This effect was transient and was not observed at any other time point tested.

Figure 5.

Intestinal IGFBP-4 mRNA expression is increased by GLP-2. A, Relative expression of IGFBP-4 mRNA in mucosal scrapes extracted from control mice treated with 0.1-μg/g hGly2-GLP-2 (GLP-2) or vehicle (PBS) for 10 days; n = 9–12. The adjacent bar graph indicates relative mRNA expression in whole jejunum of control mice; n = 4. B, Relative expression of IGFBP-4 mRNA extracted from FRIC cultures treated with 10⁻⁸M GLP-2 or vehicle for 0.5–24 hours; n = 4–12 per group. Open bar indicates vehicle treatment, and closed bar indicates GLP-2 treatment. *, P < .05 vs paired control.

Discussion

The mechanisms underlying the intestinotropic actions of GLP-2 are highly complex due to the localization of the GLP-2Rs to specialized cells which do not include the crypt cells involved in growth of the small intestine (7–10). Our previous studies have demonstrated that GLP-2-induced intestinal growth and proliferation occurs in an IGF-I- and IE-IGF-IR-dependent manner (3, 12, 13). By using both loss-of-function and gain-of-function murine models, as well as fetal rat intestinal cells in culture, the results of the present study extend this signaling pathway to include IGFBP-4.

The bioactivity of IGF-1 is known to be regulated by IGFBPs, of which only isoforms 3 through 7 were detected in murine whole-thickness jejunum and mucosal scrapes. Although IGFBP-3, IGFBP-4, and IGFBP-5 have been reported to be the most highly expressed in the gut (20), IGFBP-3 and IGFBP-5 are not essential for the growth effects of GLP-2 (14). The physiological role of IGFBP-4 is complex, with reports of both stimulatory and inhibitory effects on the activity of IGF-1 and IGF-II (22, 23). The effects of IGFBP-4 on the small intestine are largely unknown, although it has been reported to inhibit the growth of colorectal cancer cells (30). Under basal conditions in the present study (ie, vehicle-injected mice), lack of IGFBP-4 resulted in enhanced intestinal weight, such that the small intestinal weight was increased relative to both body weight and intestinal length. The latter finding suggests that IGFBP-4 inhibits basal growth of the intestine at the level of the mucosa, consistent with the observation of enhanced crypt cell proliferation in the KO mice in absence of any change in the thickness of the muscularis. In contrast, IGFBP-4 null mice displayed an impaired growth response to GLP-2, with a blunted increase in small intestinal weight normalized to either body weight or intestinal length, an impaired increase in crypt-villus height, and a completely abrogated proliferative response as detected by changes in Ki-67 expression. These results demonstrate that IGFBP-4 normally promotes the growth effects of GLP-2, presumably by modulating the bioavailability of IGF-1 and its downstream activation of the intestinal epithelial IGF1-R (3, 12, 13). A summary of the proposed mechanism of action of GLP-2 to promote intestinal growth is shown in Figure 6. However, because IGFBP-4 can also have IGF-independent effects, inhibiting cWnt (canonical Wingless-related integration site) signaling in cardiac smooth muscle by directly interacting with the Frizzled-8 receptor (31), the absolute requirement for IGF-1 in the GLP-2-related actions of IGFBP-4 remain to be determined. Finally, preliminary studies suggested that the impaired growth response to GLP-2 in the IGFBP-4 null mice could be overcome by raising the GLP-2 dose 100-fold (from 0.1 to 10 μg/g), indicating that, with a sufficiently high dose of GLP-2, the further enhancement of IGF-1 levels and/or other growth factors are able to compensate for the absence of IGFBP-4. Such mediators may include the epidermal growth factor receptor pathway previously shown to play a role in the tropic response of the small intestine to GLP-2 treatment (11, 21).

Figure 6.

Schematic representation showing proposed role of IGFBP-4 in GLP-2-stimulated intestinal growth. GLP-2, once released into the circulation from the intestinal L-cells, binds to GLP-2Rs expressed by intestinal subepithelial myofibroblast (ISEMF) cells leading to release of IGF-I. The bioactivity of this IGF-1 is enhanced by the presence of locally expressed IGFBP-4, which increases IGF-IR activation and consequent stimulation of intestinal epithelial cell (IEC) proliferation within the intestinal crypt niche. Conversely, IGFBP-4 plays an inhibitory role in basal intestinal growth, and lack of the IGFBP-4 protease, PAPP-A, which leads to increased circulating IGFBP-4 levels, results in a profound reduction in overall intestinal weight.

Interestingly, IGFBP-4 does not form a ternary complex with acid-labile subunit, which is known to limit the ability of both IGFBP-3 and IGFBP-5 to cross the endothelial layer (16); IGFBP-4 may therefore more easily diffuse from the circulation into the pericryptal space. However, the results of this study suggest that IGFBP-4 promotes GLP-2-stimulated intestinal growth in a local/paracrine manner. Specifically, the PAPP-A KO mice demonstrated a preserved GLP-2 growth response with respect to both intestinal mass and crypt-to-villus height. Because the PAPP-A KO mice are a gain-of-function model, with increased circulating levels of IGFBP-4, this suggests that circulating IGFBP-4 does not play a role in the GLP-2-stimulated intestinal growth. However, because the PAPP-A KO small intestines were significantly smaller than found in the WT mice with respect to small intestinal weight per unit length, and were notably more fragile upon handling, this finding reinforces the observation that IGFBP-4 is important for maintenance of basal small intestinal mucosal growth and suggests that circulating IGFBP-4 contributes to these actions.

Consistent with the possibility of interactions between GLP-2 and IGFBP-4, increased IGFBP-4 mRNA transcript levels were detected in mucosal scrapes from control mice treated with GLP-2 chronically for 10 days followed by an acute injection 3 hours before tissue collection. Stimulation of IGFBP-4 mRNA expression by GLP-2 was also observed in FRIC cultures after acute treatment (2 h) with GLP-2. Because FRIC cultures responded to GLP-2 in a transient manner, it seems likely that the increase in the IGFBP-4 mRNA expression seen in the mucosal scrapes of the control animals is also a transient response to the acute administration of GLP-2. Interestingly, we have previously demonstrated that the transcription factor, sex determining region Y-box 9, is a downstream target in the jejunum that is enhanced by acute (90 min) treatment of mice with GLP-2 (32). Furthermore, sex determining region Y-box 9 has been shown to directly bind to the IGFBP-4 promoter, thereby enhancing IGFBP-4 gene expression (30). Collectively, these findings provide additional support for a role of IGFBP-4 in GLP-2-stimulated intestinal growth. However, the cellular source of this IGFBP-4 currently remains unclear. Given the relatively higher levels of IGFBP-4 observed in the whole intestine as compared with the mucosa alone, it also remains possible that this binding protein is derived from cells that are not directly contained within the crypt compartment, such as the submucosa or the muscularis. Such a source of IGFBP-4 would mandate an indirect mechanism of action of GLP-2, through the GLP-2R expressed on the intestinal subepithelial myofibroblast cells, the enteric neurons or the enteroendocrine cells, to release other factors such as ErbB ligands or keratinocyte growth factor, for example (8–11), that could then stimulate IGFBP-4.

Although the present study demonstrates a role for IGFBP-4 in the signaling mechanism downstream of the GLP-2R (Figure 6), the exact mechanism whereby IGFBP-4 enhances GLP-2-stimulated IGF signaling remains to be elucidated. Studies with IGFBP-5 have shown that this protein can interact directly with the extracellular matrix to sequester and potentiate the effect of IGF signaling (33). Thus, once released and bound to IGF-I, IGFBP-4 may be interacting with either the extracellular matrix close to the intestinal epithelial cells or with the epithelial membrane itself to enhance IGF-1 effects, as for other IGFBPs (34), although this has not been described to date for IGFBP-4. It must also be recognized that circulating and locally expressed IGFBP-4 may differ in their glycosylation status (35), resulting in differences in their ability to interact with the cellular membrane and/or extracellular matrix as well as to whether IGF-independent pathways can be activated (17). Finally, it must be noted that, although IGF-1 is believed to be the predominant mediator responsible for the GLP-2-stimulated growth response (12), IGF-II has also been implicated in playing a small role, with IGF-II^−p mice displaying a significant but reduced response to GLP-2 treatment (12). Because IGF-II is also capable of binding to IGFBP-4 (25, 34) and stimulating the IGF-IR (36), it cannot be discounted that it is lack of IGF-IR activation by both IGF-1 and IGF-II that is responsible for the impaired GLP-2 growth response in the IGFBP-4 KO mice.

In conclusion, the results of this study demonstrate not only that circulating IGFBP-4 plays an inhibitory role in basal intestinal growth but also that locally produced intestinal IGFBP-4 is integral for GLP-2-stimulated intestinal growth (Figure 6). These findings provide insight into the tightly regulated molecular mechanisms underlying the intestinotropic effects of GLP-2 and may therefore have therapeutic implications for new targets for use in patients with short bowel syndrome. Further studies will be required to elucidate the signaling mechanisms that underlie the intestinal IGFBP-4 response to GLP-2 as well as the effects of IGFBP-4 to enhance IGF signaling in the gut.