Coordinated Control of Fetal Gastric Epithelial Functions by Insulin ...

0013-7227/01/$03.00/0 Endocrinology Copyright © 2001 by The Endocrine Society

Vol. 142, No. 5 Printed in U.S.A.

Coordinated Control of Fetal Gastric Epithelial Functions by Insulin-Like Growth Factors and Their Binding Proteins* ERIC TREMBLAY, PIERRE CHAILLER,

AND

´ NARD DANIEL ME

Canadian Institutes of Health Research Group on the Functional Development and Physiopathology of the Digestive Tract, Department of Anatomy and Cell Biology, Faculty of Medicine, Universite´ de Sherbrooke, Sherbrooke, Que´bec, Canada J1H 5N4 ABSTRACT The influence of insulin-like growth factors (IGFs) and their binding proteins (IGFBPs) on human gastric functions are unknown. This study was undertaken to evaluate the ability of fetal gastric mucosa to produce IGFBPs and to test the effects of IGF-I, IGF-II, and synthetic truncated IGFs that do not interact with IGFBPs on epithelial cell proliferation and glandular zymogenic function. Western blots, Far Western blots, and immunohistochemistry were performed to characterize the expression of IGFBP-1 to -6 and IGF-I receptor. The effects of growth factors on DNA synthesis and lipase and pepsin activities were determined in gastric explants maintained in serumfree organ culture. All gastric epithelial cells expressed the IGF-I receptor. IGFBP-2 to -6 were produced endogenously, and they were differentially localized along the foveolus-gland axis and modulated in culture. Exogenous IGF-I and IGF-II were able to reduce lipase

T

HE HUMAN gastric mucosa develops very early during fetal life (starting ⬃9 weeks), and adult-type compartmentalization of fundic (oxyntic) glands involved in mucus, acid, and zymogen secretion becomes rapidly established (1–3). Mucous epithelial cells differentiate from mitotic progenitors during their migration from the isthmus toward the foveolus and the surface, whereas parietal, endocrine, and chief zymogenic cells appear in the forming glands, similar to those in normal renewing adult mucosa (2, 4). Moreover, recent data argue in favor of a precocious appearance of gastric secretory functions during ontogeny. We have reported ultrastructural and biochemical evidence of significant mucus/glycoprotein production (5) and revealed immunoreactive fundic pepsinogen-5 (Pg5) and human gastric lipase (HGL) as well as pepsin and lipase enzyme activities around 10 –13 weeks (6, 7). In fact, coexpression of Pg5 and HGL in chief cells in the human fetus (7) and adult (8) represents a unique feature, a gastric lipase enzyme that is absent in rodents and distinctively localized in other mammals (9). Although a comprehensive view of the human

Received September 27, 2000. Address all correspondence and requests for reprints to: Daniel Meńard, Ph.D., De´partement d’Anatomie et de Biologie Cellulaire, Faculte´ de Me´decine, Universite´ de Sherbrooke, 3001 12e avenue N, Sherbrooke, Que´bec, Canada J1H 5N4. E-mail: [email protected]. * This work was supported by the Canadian Institutes of Health Research (to D.M.). Preliminary results were presented at the 101st Annual Meeting of the American Gastroenterological Association, San Diego, CA, and have been published in abstract form (Gastroenterology 118:A555, 2000).

activity without affecting pepsin, whereas they exerted different effects on cellular proliferation: IGF-I was stimulatory and IGF-II had no influence. Illustrating the complex regulatory effect that IGFBPs exert on IGF bioactivity, both truncated IGF-I and IGF-II stimulated DNA synthesis more than IGF-I. Moreover, the striking difference in mitogenic activity between truncated and native forms of IGF-II probably reflects the abundance of IGFBP-2 and IGFBP-6, two IGF-II carriers, in the foveolus/neck region, including the proliferative compartment. This study provides new evidence for the involvement of an intragastric IGF/IGFBP system in the fine regulation of epithelial cell division and also in the control of zymogen synthesis. Moreover, the specific influence of IGF-II as a mitogen appears to be tightly regulated by IGFBP isoforms preferentially associated with this growth factor and proliferative cells. (Endocrinology 142: 1795–1803, 2001)

gastric epithelial functions involved in restitution and cancer is emerging, our actual knowledge concerning the factors involved in the control of cell renewal and zymogen expression/synthesis in chief cells is only fragmentary. It is also of prime importance to delineate the specific mechanisms regulating the expression of HGL in fetal, neonatal, and adult gastric tissue, as initial digestion of dietary fat in stomach of infants and adults seems to be a prerequisite for efficient intestinal lipolysis (9, 10). HGL even assumes a more crucial role in physiological (term and preterm infants) and pathological conditions associated with pancreatic insufficiency and a low level of pancreatic lipase (8). Several lines of evidence suggest that insulin-like growth factors (IGF-I and IGF-II) and insulin may be involved in the development and maintenance of gastric functions. These factors are widely expressed in embryonic and adult tissues (11), and null mutations of IGF-I, IGF-II, or IGF-I receptor (IGF-I-R) genes cause retardation of fetal growth in mice (12). The IGF-I-R (also termed type I IGF-R) and insulin receptor (insulin-R) are present at the epithelial level in all segments of the gastrointestinal tract in rodents (13–15), and both ligands are mitogenic for primary cultures of canine gastric epithelial cells (16). It is noteworthy, however, that IGF-II is the most abundant insulin-related substance detected in human blood cord vessels (17) and in culture medium of human gut-derived epithelial cell lines (18 –20). IGFs are additionally complexed in blood and extracellular fluids to high affinity binding proteins (IGFBP-1 to IGFBP-6), which display sequence homology to a new group of IGFBP-related proteins,

1795

1796

IGFs/IGFBPs AND GASTRIC EPITHELIUM

but not the IGF-I-R (11, 21, 22). IGFBP-1 to -6 proteins specifically modulate the bioactivity of IGFs, and their respective affinities for each ligand are known to vary (23, 24). Finally, it was reported that human fetal mesenchymal cells synthesize IGF messenger RNA (mRNA) (25), whereas gastric mucosal cells express IGFBP-2 to -6 mRNAs (26). However, whether the mucosa is able to translate these IGFBP transcripts and to secrete the different IGFBPs is unknown, and whether IGFs are able to modulate human gastric epithelial cell proliferation and specific digestive functions independently of IGFBPs remains to be investigated. A serum-free organ culture methodology has been recently applied to human fetal gastric mucosa in our laboratory (5). Compared with strategies based on short-term culture of adult biopsies, this technique offers many advantages for investigating the roles of growth factors in gastric physiology such as a prolonged survival time and the maintenance of tissue under completely defined conditions. It enabled us to demonstrate that epidermal growth factor and transforming growth factor-␣ exert a direct influence on epithelial cell proliferation and HGL expression in chief cells (27, 28). In our ongoing effort to identify the regulators of cell proliferation and zymogen synthesis in human stomach, the present study was undertaken 1) to establish the cellular and glandular localizations of IGFBPs, 2) to investigate their secretion patterns, and 3) to verify the specific influence of IGF-I and IGF-II on cell proliferation and digestive functions (Pg5 and HGL) by comparing native factors and truncated IGFs [R3IGF-I; Del(1– 6)IGF-II], which do not interact with IGFBPs. Materials and Methods Specimens Tissues from 51 fetuses, aged 15–20 weeks [postfertilization ages estimated according to Streeter (29)], were obtained from normal elective pregnancy terminations. Studies were approved by the institutional human subject review board, and no tissue was collected from cases associated with known fetal abnormality or fetal death. The stomach was immersed in Leibovitz L-15 culture medium (room temperature) containing 40 ␮g/ml nystatin and gentamicin (Life Technologies, Inc. Life Technologies, Inc., Burlington, Ontario, Canada) and prepared within 30 min. For immunofluorescence studies, specimens were rinsed, embedded in OCT (Tissue-Tek, Miles Laboratories, Elkhart, IN), and frozen in liquid nitrogen.

Serum-free organ culture Gastric tissue was prepared as previously described (5). Briefly, the glandular mucosa (body/fundus) was cut into 5 ⫻ 5-mm2 explants and maintained in organ culture dishes (Falcon Plastics, Los Angeles, CA) at the interface of 95% air-5% CO2 gas mixture and culture medium for up to 5 days (37 C). Medium was renewed after 1 day and every 2 days thereafter. Human recombinant IGF-I, IGF-II (Collaborative Biomedicals, Bedford, MA), or insulin (Novo Nordisk Canada, Mississauga, Ontario, Canada) were added at respective concentrations of 50 –100 ng/ml (6.5–13 nm), 100 –200 ng/ml (13–26 nm), and 30 –300 ␮U/ml (1.2–12 ng/ml or 0.2–1.6 nm). These dosages correspond to circulating levels measured in fetal cord blood in normal term and preterm infants (17, 30). Two truncated IGF peptides lacking the IGFBP-binding domain (purchased from Upstate Biotechnology, Inc., Lake Placid, NY), namely R3IGF-I (50 –100 ng/ml) and Del(1– 6)IGF-II (100 –200 ng/ml), were also tested to evaluate the modulatory effect of endogenous IGFBPs on IGFinduced responses.

Endo • 2001 Vol. 142 • No. 5

Antibodies The following rabbit or mouse primary antibodies raised against human proteins were shown previously to be highly specific: clone ␣IR3 for IGF-I-R (31) (Oncogene Science, Inc., Cambridge, MA), clone cII 25.3 for insulin-R (32) (Oncogene Science, Inc.), polyclonal antibodies to IGFBP-1 to -5 (33, 34) (Upstate Biotechnology, Inc.), and polyclonal antibody to IGFBP-6 (35) (Austral Biologicals, San Diego, CA). The latter probe is recommended for enzyme-linked immunosorbent assay and immunohistochemistry, and it is known to cross-react with other IGFBPs in Western assay conditions.

Indirect immunofluorescence Detection of tyrosine kinase IGF-I-R and IGFBP isoforms on tissue cryosections was performed as described previously (27). After fixation in either 1% formaldehyde diluted in 100 mm sodium phosphate buffer, pH 7.4 (45 min at 4 C), or acetone/chloroform (1:1; 5 min at ⫺10 C), specimens were incubated in 100 mm glycine in PBS if quenching of aldehyde residues was necessary; blocked in Blotto, BSA, or fish gelatin solutions; and then incubated for 1 h at room temperature with primary antibodies diluted in 2% BSA-PBS. Fluorescein-conjugated secondary antibodies were then added for 45 min. Control experiments were performed by omitting or replacing the primary antibody with the appropriate nonimmune serum. Primary antibodies were used at the following dilutions: IGF-I-R, insulin-R, IGFBP-1, IGFBP-3, IGFBP-4, and IGFBP-5, 1:100; IGFBP-2, 1:1000; and IGFBP-6, 1:25. Antimouse (1:30) and antirabbit (1:50) secondary antibodies were obtained from Roche Molecular Biochemicals Canada (Laval, Que´bec, Canada).

Western and Far Western immunoblotting SDS-PAGE was performed as described previously for fetal stomach (28) on 12% and 15% acrylamide gels. Total proteins from gastric explants were extracted in 2 ⫻ sample buffer [Tris-HCl (pH 6.8), 4% SDS, and 2% ␤-mercaptoethanol]. Proteins in culture medium were concentrated approximately 30- to 50-fold by ultrafiltration using Centricon-10 concentrators (Amicon Canada, Oakville, Ontario, Canada), then mixed with 2 ⫻ sample buffer. Aliquots of tissue proteins (180 –200 ␮g) and medium proteins (80 –100 ␮g) assayed by the method of Lowry (36) were separated, transferred, incubated with primary antibodies after blocking in 0.2% highly purified casein, and then processed with the WesternLight Plus Chemiluminescent Detection System (Tropix, Bedford, MA). Autoradiograms exposed in a linear range were quantified by densitometric analysis with an LKB XL Ultroscan (Pharmacia Biotech, Baie d’Urfe´, Que´bec, Canada), and signals were normalized to a keratin-18 control. Antibodies were used at the following dilutions: IGFBP-1, 1:1000; IGFBP-2, 1:2000; IGFBP-3, 1:800; IGFBP-4, 1:1000; IGFBP-5, 1:750; IGFBP-6, 1:100 –1000; and keratin-18 (monoclonal CY-90, Sigma, St. Louis, MO), 1:5000. For Far Western blots, SDS-PAGE and immunodetection were performed as described above, except that all of these steps were carried out under nonreducing conditions to allow the binding of IGF to IGFBPs (37). After electrotransfer, membranes were incubated with a relatively high concentration of human IGF-I (200 ng/ml) and probed with a monoclonal antibody against this peptide (2 ␮g/ml; from Upstate Biotechnology, Inc.).

Tritiated thymidine incorporation To determine the rate and site of DNA synthesis, 2 ␮Ci [3H]thymidine (80 Ci/mmol; Amersham Pharmacia Biotech Canada, Oakville, Ontario, Canada) were added per ml medium during the last 6 h of culture (2, 27). The level of radiopercursor incorporated into trichloroacetic acidprecipitable material was expressed as disintegrations per min/mg tissue and reported as stimulation percentage vs. the control value. For radioautography of incorporation sites, some explants were fixed in 2.8% glutaraldehyde in 0.1 mol/L cacodylate for 16 –24 h at 4 C, postfixed in 2% osmium tetroxide in cacodylate for 30 min, dehydrated, and embedded in epon resin. One-micron sections were mounted on glass slides, stained with aldehyde fuschin, dipped in Kodak NTB2 emulsion (Eastman Kodak Co., Rochester, NY), and then exposed in the dark for 6 weeks. The [3H]thymidine incorporation labeling index was estab-

IGFs/IGFBPs AND GASTRIC EPITHELIUM lished at the epithelial level and calculated as the percentage of labeled nuclei (5 or more grains overlying the nucleus) over the total number of nuclei counted in the entire epithelium. For each fetus, 2–3 explants were processed, and 1000 nuclei were counted/explant.

1797

detection of affinity-bound ligand (Fig. 2A). In three different specimens, seven bands were visualized in the 28 – 47 kDa weight range that could correspond to previously characterized IGFBP proteins: IGFBP-1, 32 kDa;

Lipase and pepsin enzymatic assays Lipolytic activity was measured using a long-chain triglyceride substrate, i.e. glycerol tri-[14C]oleate (Amersham Pharmacia Biotech Canada; SA, 59 mCi/mmol) as detailed previously (6). Released free [14C]oleic acid was separated by liquid-liquid partition in methanol/ chloroform/heptane, collected in the supernatant upon precipitation with carbonate-borate buffer, pH 10.5, and quantitated by liquid scintillation spectrometry. Activity was expressed as nanomoles of FFA produced per min/mg tissue. Pepsin activity was assayed by a modification of the method of Anson and Mirsky (38), using dialyzed 2% hemoglobin (H-2625, Sigma) as substrate. The free amino acids generated by pepsin were measured in the supernatant by spectrometry (280 nm) using an l-tyrosine standard. Data were expressed in pepsin units, i.e. ␮mol of tyrosine-containing peptides released per 10 min.

Statistics Numeric values are reported as the mean ⫾ sd. Statistical significance of differences between experimental conditions was established at 95% and determined by ANOVA followed by Student’s t test when significance was indicated.

Results Localization of IGF-I-R and insulin-R

Immunofluorescence staining of IGF-I-R was performed on 16- and 20-week-old fetal gastric mucosa. IGF-I-R were present on the basolateral membranes of surface epithelial cells as well as those of all epithelial cells of the foveolusgland axis (Fig. 1, A and B). A similar distribution of insulin-R was visualized (not shown), and no significant immunostaining was observed in all cases when the primary antibody was omitted or replaced by the appropriate nonimmune serum (Fig. 1C). Far Western and Western immunoblotting

In a first step, gastric mucosal IGFBPs were detected using an IGF-I binding procedure followed by immuno-

FIG. 2. Detection of tissue-derived IGFBPs by Far Western blot technique (immunostaining of affinity-bound IGF-I; see Materials and Methods). A, Total proteins from fetal stomach tissue were resolved on 12% acrylamide gels, and seven IGFBP bands (from 28-kDa to 47-kDa), named a to g, were visualized in specimens of three age categories (15, 17, and 20 weeks; the last is shown). B, On 15% acrylamide gels, a low molecular mass IGFBP of 22 kDa (named x) was revealed in the culture medium (M) below the 28-kDa IGFBP (named a). The 22-kDa form was absent in the corresponding tissue homogenate (T).

FIG. 1. Expression and distribution of IGF-I-R in the developing human gastric mucosa. Indirect immunofluorescence micrographs of cryosections of corpus regions of fetal stomach at 16 (A) and 20 (B) weeks gestation stained with monoclonal ␣IR3 antibody. Receptors were localized at the basolateral membranes of all epithelial cells along the surface/foveolus/gland axis, and no significant staining was observed using nonimmune serum as a control (C). Magnification, A–C, ⫻190. Bar, 100 ␮m.

1798


IGFBP-2, 34 kDa; IGFBP-3 doublet, 39 – 47 kDa; IGFBP-4, 28 kDa; IGFBP-5, 31 kDa; and IGFBP-6, 36 kDa. No quantification was performed, however, because immunoreactive signals are proportional not only to the protein level, but also to the affinity of the respective IGFBPs for exogenous IGF-I in this assay. Therefore, specific IGFBP isoforms were identified by classical Western blot using polyclonal antibodies in either uncultured tissue or explants cultured for 24 h and their corresponding media. AntiIGFBP-6 was tested at various dilutions and was judged inappropriate for Western blot analysis; it reacted with a 36-kDa protein (a diffuse band, possibly corresponding to IGFBP-6) and with other forms (34 kDa, IGFBP-2; results not shown), as previously reported for intestinal cells (35). In these assays however, IGFBP-1 to -5 proteins were successfully detected (Fig. 3, A–E), and immunoreactivity for IGFBP-3 was particularly intense, suggesting that the latter carrier is the most abundant in intact gastric tissue, followed by IGFBP-1, IGFBP-2, and IGFBP-5 (compare left lanes). The staining intensities may vary slightly according to the labeling properties of each probe. In organ culture, the total amounts (tissue plus medium) of IGFBP proteins all increased, except for IGFBP-1 (Fig. 3F). Indeed, IGFBP-2 to -5 were actively synthesized during the first day of culture, as their total levels increased by up to 3-fold compared with that in uncultured tissues, as seen for IGFBP-2. Although immunoreactive IGFBP-1 to -4 were efficiently released in the culture fluid, no IGFBP-5 was detected. This suggests that either IGFBP-5 is delivered by a different cellular mechanism, or it is cleaved into proteolytic fragments that are not recognized by the antibody used. In this regard, previous studies have demonstrated that IGFBP-5 is very sensitive to the actions of serine proteases and is only present in cell culture medium in the form of 20-, 22-, or 24-kDa fragments (39 – 41). Thus, a second set of Far Western ligand blot experiments was performed on 15% acrylamide gels to verify whether IGFBP fragments of lower molecular mass were present in the culture medium of gastric explants (Fig. 2B). One binding protein of approximately 22 kDa that was not found in the corresponding tissue homogenate was successfully visualized. Tissue distribution of IGFBPs

We then characterized in 12- to 20-week specimens the cellular localization of IGFBP-1 to -6 proteins along the functional foveolus-gland axis with the same antibodies (Fig. 4). IGFBP-1 immunostaining was visualized at the level of epithelium as well as in several elements of the lamina propria (mesenchyme and blood vessels; Fig. 4A). No evidence of restricted synthesis site was found. By contrast, other isoforms were preferentially associated with epithelial cells. IGFBP-3, IGFBP-4, and IGFBP-5 (Fig. 4, B–D) were detected in the entire epithelium, generally with increased reactivity in surface epithelial cells. The expression of IGFBP-2 protein was more rigorously compartmentalized. An intense cytoplasmic and basolateral staining was seen in epithelial cells of the foveolus/neck region, including the proliferative compartment, whereas the base of glands was negative. This

Endo • 2001 Vol. 142 • No. 5

FIG. 3. Detection of tissue-derived IGFBPs by Western blot technique using isoform-specific antibodies. Immunolabeling was performed in uncultured gastric tissue (left lane), explants incubated for 24 h (center lane) and their corresponding media (right lane): A, IGFBP-1; B, IGFBP-2; C, IGFBP-3 doublet; D, IGFBP-4; E, IGFBP-5. Densitometric analyses (F) comparing the relative amounts of each BP protein in cultured tissue and media vs. intact tissue (To ⫽ 100%) revealed that IGFBP-1 was not synthesized de novo as opposed to other isoforms and that the synthesis/secretion of IGFBP-2 were particularly intense in culture.

observation was made in either early forming (Fig. 4E) or maturing units (Fig. 4F). The distribution of IGFBP-6, consistent with endogenous expression in epithelial cells, also varied greatly among the compartments. Compared with superficial epithelium, the bases of primitive (Fig. 4G) and differentiated glands (Fig. 4H) were mildly positive, and staining was redistributed to the apical cell region. Again, no significant immunoreactivity was observed in controls (not shown).


1799

FIG. 4. Indirect immunofluorescence of IGFBPs. The results shown are representative of three to five experiments performed in 11- to 20-week gestation specimens at the level of body/fundus regions. IGFBP-1 reactivity (A) was detected in many cell types. IGFBP-3 (B), IGFBP-4 (C), and IGFBP-5 (D) were visualized in epithelial cells along the foveolus-gland unit. IGFBP-2 was restricted to the surface and superior half of glands (delineated in white) at 13 weeks (E) and 20 weeks (F). IGFBP-6 was also more abundant in the same compartments at 11 weeks (G) and 20 weeks (H). Arrows indicate the bases of glands where staining is less intense and redistributed to the apical cell domain. Magnification: A–D, ⫻75 (bar, 200 ␮m); E–G, ⫻190 (bar, 100 ␮m); H, ⫻130 (bar, 100 ␮m).

DNA synthesis

Explants of fundic-type mucosa were cultured for 24 h in the presence or absence of IGF-I, IGF-II, and insulin, and the incorporation of [3H]thymidine into total DNA was determined during the last 6 h (Fig. 5A). At the 50 ng/ml concentration, IGF-I increased the rate of total DNA synthesis compared with control value in 10 of 13 cultures (mean stimulation, 45.19 ⫾ 12.70%; P ⬍ 0.01). This effect was al-

ready maximal (similar at 100 ng/ml) and was maintained after 48 h (results not shown). By contrast, neither IGF-II nor insulin significantly stimulated [3H]thymidine incorporation after the same time intervals (results for 100 ng and 30 ␮U/ml dosages are illustrated, respectively). Explants from three different specimens were processed for radioautography to determine the sites of precursor incorporation at the level of the epithelium. The labeled nuclei were found mainly in the

1800


Endo • 2001 Vol. 142 • No. 5

FIG. 5. [3H]Thymidine incorporation into DNA during the last 6 h of 24-h cultures of gastric explants, aged 15–20 weeks gestation. A, The effects of growth factors IGF-I, IGF-II, and insulin (50 ng/ ml, 100 ng/ml, and 30 ␮U/ml dosages are shown, respectively). B, Comparison of IGFs with their respective synthetic analogs, R3IGF-I (50 ng/ml) and Del(1– 6)IGF-II (100 ng/ml). Incorporation is calculated as disintegrations per min/␮g DNA and is illustrated as the percentage of increase or decrease relative to its own control value. Each bar represents the result for a single experiment, and the mean ⫾ SD of 5–13 experiments (numbers in parentheses) are given in the text.

foveolus/neck region of gastric glands, as previously reported (2). IGF-I supplementation did not alter their overall morphology and induced a significant increase (42%) in the epithelial labeling index compared with that in untreated explants (control, 17.9 ⫾ 2.0% of total cells labeled; IGF-I, 25.6 ⫾ 2.9%; P ⬍ 0.006). In a second set of experiments, the effects of native IGFs were compared with those of truncated analogs that do not interact with IGFBPs (Fig. 5B). R3IGF-I added at the same concentration as native IGF-I, i.e. 50 ng/ ml, induced a similar and stronger stimulation of [3H]thymidine incorporation into total DNA (R3IGF-I, 68.22 ⫾ 14.23%; IGF-I, 44.68 ⫾ 14.89%; P ⬍ 0.04). Addition of 100 ng/ml Del(1– 6)IGF-II caused a strong 1-fold increase in the incorporation level (114.04 ⫾ 11.60%; P ⬍ 0.0005), whereas native IGF-II remained without effect. Lipase and pepsin activities

Explants were maintained for up to 5 days with or without IGF-I, IGF-II, insulin, or truncated IGF peptides, and enzymatic activities were determined in tissue homogenates (Fig. 6) as well as in culture medium. For lipase, although the amount of enzyme secreted did not vary (not shown), there was a significant decrease in tissue activity in explants treated with IGFs [Fig. 6A; 50 ng/ml IGF-I, ⫺31,6% (P ⬍ 0.027 vs. control); 100 ng/ml IGF-II, ⫺42,3% (P ⬍ 0.022 vs. control)]. In comparison, both truncated R3IGF-I and Del(1– 6)IGF-II decreased tissue levels of lipase activity slightly more than IGFs [Fig. 6A; R3IGF-I (⫺57.9%) vs. IGF-I (⫺31.6%), P ⬍ 0.029; Del(1– 6)IGF-II (⫺62.9%) vs. IGF-II (⫺42.3%), P ⬍ 0.08, marginally significant]. As opposed to lipase, pepsin activity resulting from activation of fundic-type pepsinogen (Pg5) at acid pH was not modulated by these factors (Fig. 6B), and insulin did not influence lipase or pepsin activities during the same culture period (data not shown). Discussion

Insulin-related factors are recognized as potent mitogens for a wide variety of cell types, including epithelial cells (11). They commonly act as progression factors in culture, i.e. by stimulating rapid cell division in the presence of other hor-

mones/growth factors required for basal proliferation and survival. The biological action of such progression factors seems particularly relevant in the context of processes requiring a rapid reentry in mitosis, such as compensatory growth, regeneration (reepithelialization), or simply induction of rapid cell renewal by physiological stimuli. Recent studies (cited in Ref. 42) reinforce the view that IGFs represent autocrine/paracrine inducers of epithelial cell growth involved in intestinal adaptation after injury or resection. The current investigation establishes that immunoreactive IGF-I-R and insulin-R are expressed at the level of basolateral membranes of human gastric epithelial cells, corroborating data mainly obtained through binding studies in rodents (13–15) and the intestinal HT-29-D4-Gal cell line (35). Both surface and glandular epithelia were found to be positive, suggesting that all epithelial cells, including mitotic precursors, would be competent to respond to IGF-I, IGF-II, and/or insulin. Organ culture experiments indeed demonstrate that IGF-I is able to stimulate epithelial cell proliferation in intact gastric tissues whereas both IGFs reduce HGL activity in fetal chief cells, similar to epidermal growth factor and transforming growth factor-␣ in the same model (27, 28). The mitogenic effects of IGF-I on normal human gastric epithelial cells would be consistent with those reported with monolayer cultures of human gastric cancer cells (43). Also, this would agree with the enhanced growth of the small bowel observed in transgenic mice overexpressing IGF-I (44). However, the lack of IGF-II mitogenic activity contrasts with its stimulatory effect on enterocyte-like Caco-2 cells constitutively expressing the IGF-II transgene (45). The current study also shows that both IGFs repress HGL activity without affecting pepsin, an observation that further supports the existence of a functional uncoupling between the expression of the two gastric zymogens either during in utero development (6) or in organ culture (5, 27). The fall in HGL activity induced by IGFs is reminiscent of the down-regulation of intestinal sucrase-isomaltase activity by IGF-II (45). To assess whether IGFBPs have the capacity to modulate IGF biological actions either negatively or positively, we extended previous analyses establishing the presence of


1801

FIG. 6. Lipase (A) and pepsin (B) activities in gastric tissue cultured 5 days without (C, control) or with IGF-I (50 ng/ml), IGF-II (100 ng/ml), and their respective synthetic analogs, R3IGF-I and Del(1– 6)IGF-II. Values represent the mean ⫾ SD of five separate and independent cultures. Statistically significant differences between truncated (f) or native (䡺) IGFs compared with control, P ⬍ 0.027 (*). Statistically significant difference between truncated and native IGF-I, P ⬍ 0.029. Marginally significant difference between truncated and native IGF-II, P ⬍ 0.08.

IGFBP-2, IGFBP-3, IGFBP-4, IGFBP-5, and IGFBP-6 mRNAs in human fetal stomach (26). Indeed, the intrinsic capacity of human gastric epithelial cells to translate these mRNAs and to secrete the proteins remained uncertain. Far Western blotting experiments helped to identify seven IGFBPs by the affinity-bound IGF-I technique (37) that were tentatively associated to known IGFBP-1 to – 6, including the 39 – 47 kDa doublet of IGFBP-3. Western blot analyses using specific antibodies (33, 34) confirmed the presence of IGFBP-1 to -5 with their expected molecular masses in intact tissues as well as in cultured gastric explants and their corresponding media. These experiments clearly established the capacity of the human gastric mucosa to translate IGFBP mRNAs (IGFBP-1 excepted) and to efficiently secrete the various IGFBPs. Incidentally, the total amount of IGFBP-1 remained constant compared with that in uncultured gastric explants, indicating that the gastric pool of IGFBP-1 probably derives from another source, the fetal liver, as the corresponding mRNA is expressed only in this organ (26). This study also enlightens different secretion patterns. Intact IGFBP-1 to -4 proteins were all released into culture medium. IGFBP-2 was the most secreted protein, whereas a 22-kDa IGFBP form, probably corresponding to a proteolytic fragment of IGFBP-5, was detected. Keeping in mind that IGFBP-5 remains preferentially associated in vivo and in vitro with extracellular matrix proteins (46) and that IGFBP-5 fragments isolated from the medium of cultured cells exhibit a reduced, but significant, IGF binding capacity (39 – 41), it can be suggested that intact and cleaved IGFBPs together contribute to the modulation of IGF activity through specific mechanisms. According to a new hypothesis, IGFBP-5 fragments may even stimulate the

phosphorylation of a putative 420-kDa receptor and mediate IGF-independent effects (40). To determine whether these IGFBPs were strategically located in specific functional compartments along the foveolus-gland unit, immunohistochemistry was performed with specific antibodies. IGFBP-1 was revealed at the level of epithelium and all components of the lamina propria, consistent with its extragastric origin (26). IGFBP-3, IGFBP-4, and IGFBP-5 were, instead, concentrated in gastric epithelium and associated with all epithelial cell types, although with variable staining intensities. Most importantly, IGFBP-2 exhibited a restricted pattern of distribution; an intense immunoreactivity was detected in the surface mucous epithelium and the junctional proliferative zone. The basal half of forming glands, which contains zymogen- and acid-secreting populations (chief, parietal) (2, 7), was negative. Although it was visualized at the base of glands, anti-IGFBP-6 staining also increased greatly toward the surface. This study clearly illustrates that IGFBP-2 to -6 are either restricted or evenly distributed in the different functional compartments, reinforcing the concept that they are strategically located to act as autocrine/paracrine modulators of IGFs actions. This is particularly evident for IGFBP-2 and IGFBP-6, which are highly concentrated in surface mucous cells and proliferative cells, are known to bind IGF-II with a greater affinity than IGF-I, and are considered IGF-II carriers (21, 24). To establish whether these IGFBPs have the capacity to positively or negatively modulate native IGFs actions, we compared their effects with those of truncated analogs that do not interact with IGFBPs (47). The observation that the analogs were slightly more efficient than native IGFs for

1802

Endo • 2001 Vol. 142 • No. 5


reducing HGL activity at physiological concentrations (50 – 100 ng/ml) suggests that the binding proteins marginally influence the response of zymogenic chief cells to added IGFs. A similar conclusion could be drawn concerning the effects of IGF-I on cell proliferation, as synthetic R3IGF-I was slightly more potent than the native growth factor. However, the use of synthetic Del(1– 6)IGF-II revealed the existence of an important negative down-modulation exerted by IGFBPs on the mitogenic effects of IGF-II. Although the addition of native IGF-II remains without any significant effect on gastric cell proliferation (even at the 200 ng/ml dosage), the synthetic analog was very potent. Therefore, these observations clearly establish that the influence of IGF-II as a growthpromoting agent is tightly regulated by local expression and/or secretion of IGF-II carriers. The use of an organ culture system preserving the morphological integrity of gastric tissues and mucosal compartments further allows us to propose that the sequestering mechanism determining the local amounts of free and bioactive IGF-II represents the summation of two complementary events: the binding of IGF-II to IGFBPs secreted in the culture fluid as well as its binding to carriers present in the microenvironment of a given epithelial population, i.e. precursor cells (proliferative compartment) or chief cells (base of glands). In this regard, the relative abundance of IGFBP-2 and IGFBP-6 in the precursor cell environment (visualized by immunofluorescence) would explain why exogenous IGF-II does not stimulate mucosal growth in contrast to its truncated analog. Interestingly, two independent studies have shown that IGF-II, IGFBP-2, and IGFBP-6 are normally expressed in cultures of intestinal cells (48) and that their reentry in the rapid growth phase correlates with down-regulation of IGFBP-2 synthesis (49). The physiological relevance for expressing two endogenous IGFBPs may be related to their different intracellular sorting and delivery mechanisms (IGFBP-2, basolateral, vs. IGFBP-6, apical), as observed herein and reported for intestinal cells (35). Taken together, these observations suggest that a fine balance between the latter three peptides may be a critical determinant regulating the local amount of free IGF-II and the slow or rapid proliferation state of dividing populations of gastric epithelial cells. The less abundant IGF-I peptide, mainly synthesized by the fetal liver and gut stromal cells (25, 50), would represent the endocrine/paracrine component of the gastric IGF system that is regulated less rigorously than local IGF-II and is involved in basal stimulation of proliferation. In conclusion, this investigation provides new evidence for the presence and direct implication of an intragastric IGF/ IGFBP system in the fine regulation in epithelial cell division and in the control of zymogen synthesis at the level of maturing chief cells. The mitogenic influence of IGF-II appears to be tightly regulated by IGFBP-2 and IGFBP-6 isoforms preferentially associated with this growth factor and proliferative cells. A working model for future studies regarding the induction of rapid cell division (regeneration) in human gastric epithelium by progression factors would consider the involvement of locally produced IGF-II dissociated from IGFBP-2 and IGFBP-6 complexes after the action of IGFBP proteases released during injury or stress (kallikreins, cathepsins, and matrix metalloproteinases) (51). The novel finding

that the mucosal content of IGF-II specifically increases after intestinal resection (52) suggests that a common regulatory mechanism may apply to different gastrointestinal tract segments. Acknowledgments The authors thank L. Corriveau and J.-P. Lebel for technical assistance, and Drs. C. Poulin and F. Jacot, De´partement de la Sante´ Communautaire du Center Hospitalier Universitaire de Sherbrooke, for excellent cooperation in providing tissue specimens for this study.

References 1. Johnson LR 1985 Functional development of the stomach. Annu Rev Physiol 47:199 –215 2. Meńard D, Arsenault P 1990 Cell proliferation in the developing human stomach. Anat Embryol 182:509 –516 3. Montgomery RK, Mulberg AE, Grand RJ 1999 Development of the human gastrointestinal tract: twenty years of progress. Gastroenterology 116:702–731 4. Karam SM 1999 Lineage commitment and maturation of epithelial cells in the gut. Front Biosci 4:D286 –D298 5. Meńard D, Arsenault P, Monfils S 1993 Maturation of human fetal stomach in organ culture. Gastroenterology 104:492–501 6. Meńard D, Monfils S, Tremblay E 1995 Ontogeny of human gastric lipase and pepsin activities. Gastroenterology 108:1650 –1656 7. Basque JR, Meńard D 2000 Establishment of culture systems of human gastric epithelium for the study of pepsinogen and gastric lipase synthesis/secretion. Microsc Res Tech 48:293–302 8. Abrams CK, Hamosh M, Dutta SK, Hubbard VS, Hamosh P 1987 Role of nonpancreatic lipolytic activity in exocrine pancreatic insufficiency. Gastroenterology 92:125–129 9. Hamosh M 1990 Lingual and Gastric Lipases: Their Role in Fat Digestion. CRC Press, Boca Raton, pp 1–239 10. Hamosh M 1996 Digestion in the newborn. Clin Perinatol 23:191–209 11. Stewart CEH, Rotwein P 1996 Growth, differentiation, and survival: multiple physiological functions for insulin-like growth factors. Physiol Rev 76:1005–1026 12. Baker J, Liu JP, Robertson EJ, Efstratiadis A 1993 Role of insulin-like growth factors in embryonic and postnatal growth. Cell 75:83–95 13. Termanini B, Nardi RV, Finan TM, Parikh I, Korman LY 1990 Insulin-like growth factor I receptors in rabbit gastrointestinal tract. Characterization and autoradiographic localization. Gastroenterology 99:51– 60 14. Laburthe M, Rouyer-Fessard C, Gammeltoft S 1988 Receptors for insulin-like growth factors I and II in rat gastrointestinal epithelium. Am J Physiol 254:G457–G462 15. Sodoyez-Goffaux F, Sodoyez JC, De Vos CJ 1985 Insulin receptors in the gastrointestinal tract of the rat fetus: quantitative autoradiographic studies. Diabetologia 28:45–50 16. Chen MC, Lee AT, Soll AH 1991 Mitogenic response of canine fundic epithelial cells in short-term culture to transforming growth factor ␣ and insulinlike growth factor I. J Clin Invest 87:1716 –1723 17. Bennett A, Wilson DM, Liu F, Nagashima R, Rosenfeld RG, Hintz RL 1983 Levels of insulin-like growth factors I and II in human cord blood. J Clin Endocrinol Metab 57:609 – 612 18. Culouscou JM, Remacle-Bonnet M, Garrouste F, Fantini J, Marvaldi J, Pommier G 1990 Production of insulin-like growth factor II (IGF-II) and different forms of IGF-binding proteins by HT-29 human colon carcinoma cell line. J Cell Physiol 143:405– 415 19. Park JH, McCusker RH, Vanderhoof JA, Mohammadpour H, Harty RF, MacDonald RG 1992 Secretion of insulin-like growth factor II (IGF-II) and IGF-binding protein-2 by intestinal epithelial (IEC-6) cells: implications for autocrine growth regulation. Endocrinology 131:1359 –1368 20. Singh P, Dai B, Yallampalli C, Xu Z 1994 Expression of IGF-II and IGF-binding proteins by colon cancer cells in relation to growth response to IGFs. Am J Physiol 267:G608 –G617 21. Ranke MB, Elmlinger M 1997 Functional role of insulin-like growth factor binding proteins. Horm Res [Suppl 4] 48:9 –15 22. Clemmons DR 1998 Role of insulin-like growth factor binding proteins in controlling IGF actions. Mol Cell Endocrinol 140:19 –24 23. Zapf J 1995 Physiological role of the insulin-like growth factor binding proteins. Eur J Endocrinol 132:645– 654 24. Binoux M 1996 Insulin-like growth factor binding proteins (IGFBPs): physiological and clinical implications. J Pediatr Endocrinol Metab [Suppl 3] 9:285–288 25. Han VKM, Lund PK, Lee DC, D Ercole AJ 1988 Expression of somatomedin/ insulin-like growth factor messenger ribonucleic acids in the human fetus: identification, characterization, and tissue distribution. J Clin Endocrinol Metab 66:422– 429

IGFs/IGFBPs AND GASTRIC EPITHELIUM 26. Han VKM, Matsell DG, Delhanty PJD, Hill DJ, Shimasaki S, Nygard K 1996 IGF-binding protein mRNAs in the human fetus: tissue and cellular distribution of developmental expression. Horm Res 45:160 –166 27. Tremblay E, Monfils S, Meńard D 1997 Epidermal growth factor influences cell proliferation, glycoproteins synthesis and lipase activity in human fetal stomach. Gastroenterology 112:1188 –1196 28. Tremblay E, Basque JR, Rivard N, Meńard D 1999 Epidermal growth factor and transforming growth factor ␣ down-regulate human gastric lipase gene expression.Gastroenterology 116:831– 841 29. Streeter GL 1920 Weight, sitting head, head size, foot length and menstrual age of the human embryo. Contr Embryol 11:143–179 30. Spellacy WN, Buhi WC, Bradley B, Holsinger KK 1973 Maternal, fetal and amniotic levels of glucose, insulin and growth hormone. J Obstet Gynecol 41:323–331 31. Kull FC, Jacobs S, Su YF, Svobuda ME, Van Wyk JJ, Cuatrecasas P 1983 Monoclonal antibodies to receptor for insulin and somatomedin C. J Biol Chem 258:6561– 6566 32. Ganguly S, Petruzelli LM, Herrera R, Stadtmauer L, Rosen OM 1985 Insulindependent tyrosine protein kinase: cellular distribution and copurification with the insulin receptor. Curr Top Cell Regul 27:83–94 33. Hill DJ, Camacho-Hubner C, Rashid P, Strain AJ, Clemmons DR 1989 Insulin-like growth factor (IGF)-binding protein release by human fetal fibroblasts: dependency on cell density and IGF peptides. J Endocrinol 122:87–98 34. Camacho-Hubner C, Busby Jr WH, McCusker RH, Wright G, Clemmons DR 1992 Identification of the forms of insulin-like growth factor-binding proteins produced by human fibroblasts and the mechanisms that regulate their secretion. J Biol Chem 267:11949 –11956 35. Remacle-Bonnet M, Garrouste F, El Atiq F, Marvaldi J, Pommier G 1995 Cell polarity of the insulin-like growth factor system in human intestinal epithelial cells. Unique apical sorting of insulin-like growth factor binding protein-6 in differentiated human colon cancer cells. J Clin Invest 96:192–200 36. Lowry OH, Rosebrough NF, Farr AL, Randall RJ 1951 Protein measurement with the Folin phenol reagent. J Biol Chem 193:265–275 37. Hossenlopp P, Seurin D, Segovia-Quinson B, Hardouin S, Binoux M 1986 Analysis of serum insulin-like growth factor binding proteins using Western blotting: use of the method for titration of the binding proteins and competitive studies. Anal Biochem 154:138 –143 38. Anson ML, Mirsky AE 1932 The estimation of pepsin with hemoglobin. J Gen Physiol 10:342– 44 39. Nam TJ, Busby WH, Clemmons DR 1994 Human fibroblasts secrete a serine protease that cleaves insulin-like growth factor-binding protein-5. Endocrinology 135:1385–1391

1803

40. Andress DL 1998 Insulin-like growth factor-binding protein-5 (IGFBP-5) stimulates phosphorylation of the IGFBP-5 receptor. Am J Physiol 274:E744 –E750 41. Zheng B, Clarke JB, Busby WH, Duan C, Clemmons DR 1998 Insulin-like growth factor-binding protein-5 is cleaved by physiological concentrations of thrombin. Endocrinology 139:1708 –1714 42. Lund PK 1998 Molecular basis of intestinal adaptation: the role of the insulinlike growth factor system. Ann NY Acad Sci 859:18 –36 43. Guo YS, Beauchamp RD, Jin GF, Townsend Jr CM, Thompson JC 1993 Insulin-like growth factor-binding protein modulates the growth response to insulin-like growth factor 1 by human gastric cancer cells. Gastroenterology 104:1595–1604 44. Ohneda K, Ulshen MH, Randall Fuller C, D Ercole AJ, Kay Lund P 1997 Enhanced growth of small bowel in transgenic mice expressing human insulinlike growth factor I. Gastroenterology 112:444 – 454 45. Zarrilli R, Romano M, Pignata S, Casola S, Bruni CB, Acquaviva AM 1996 Constitutive insulin-like growth factor-II expression interferes with the enterocyte-like differentiation of Caco-2 cells. J Biol Chem 271:8108 – 8114 46. Jones JI, Gockerman A, Busby Jr WH, Camacho-Hubner C, Clemmons DR 1993 Extracellular matrix contains insulin-like growth factor binding protein-5: potentiation of the effects of IGF-1. J Cell Biol 121:679 – 687 47. Francis GL 1992 Novel recombinant fusion protein analogues of insulin-like growth factor (IGF)-I indicate the relative importance of IGF-binding protein and receptor binding for enhanced biological potency. J Mol Endocrinol 8:213–223 48. Zhang Y, Wick DA, Seetharam B, Dahms NM 1995 Expression of IGF-II and IGF binding proteins in differentiating human intestinal Caco-2 cells. Am J Physiol 269:E804 –E813 49. Jehle PM, Fussgaenger RD, Blum WF, Angelus NK, Hoeflich A, Wolf E, Jungwirth RJ 1999 Differential autocrine regulation of intestine epithelial cell proliferation and differentiation by insulin-like growth factor (IGF) system components. Horm Metab Res 31:97–102 50. MacDonald RG, McCuster RH, Blackwood DJ, Vanderhoof JA, Park JH 1998 Production of insulin-like growth factor I (IGF-I) and IGF-binding proteins by rat intestinal stromal cells in vitro. Cell Mol Life Sci 54:158 –166 51. Rajah R, Katz L, Nunn S, Solberg P, Beers T, Cohen P 1995 Insulin-like growth factor binding protein (IGFBP) proteases: functional regulators of cell growth. Prog Growth Factor Res 6:273–284 52. Wiren M, Adrian TE, Arnelo U, Permert J, Staab, Larsson J 1998 An increase in mucosal insulin-like growth factor II content in postresectional rat intestine suggests autocrine or paracrine growth stimulation. Scand J Gastroenterol 33:1080 –1086