Main

The biologic effects observed after insulin stimulation are pleitropic and can generally be divided into two distinct categories: mitogenic responses and metabolic responses. Regarding the gastrointestinal tract, accumulating evidence indicates that insulin determines a physiologic action on mammal intestinal growth and tissue differentiation during the fetal, neonatal, and nursing periods(1–3). Also, it is believed that insulin is a key hormone regulating the ontogenic expression of intestinal enzymes that occurs at the beginning of the rat weaning period (d 14-17 postpartum). This is attested by the observations that insulin promotes proliferation of intestinal cells in culture(4) and that parenteral administration of insulin to infant mice(1, 2) and rats(5) induces a premature and adult expression of BBM (disaccharidases), lysosomal(N-acetyl-β-glucuronidase), microsomal (sulfatase C), and cytosolic (lactate dehydrogenase) enzymes(5). These enzyme changes induced by insulin occur in the jejunum as well as in the ileum(5, 6). Oral insulin added to the diet of newborn animals produces trophic effects on ileal mucosal mass and increase ileal LPH activity(6). To the reverse, anti-IR MAb given to chicken embryos(7) delays growth and results in a mortality proportionally to the dose of antibodies administered. Despite major advances within recent years, it is not yet clear if the mitogenic and metabolic actions of insulin share a common mechanistic pathway triggered by the activation of the insulin receptor or whether they are the result of two or more diverging pathways that could even be mediated in part by binding of the hormone to IGF-I receptors, which were found to be abundant at the basolateral side of intestinal cells(8). In previous studies(9, 10), we have shown that the premature stimulation of BBM hydrolases by insulin is independent from endogenous corticosterone secretion and from hypoglycemia. After a single dose of insulin, the premature induction of SI can be detected very rapidly (within 6 h) in both differentiated villus and nondifferentiated crypt cells and is virtually inhibited by actinomycin D(10), suggesting a regulation of the SI gene by the hormone or by intracellular messengers at the transcriptional level. Shulman et al.(11) found no effect of oral insulin on LPH precursors synthesis nor on LPH mRNA levels in the ileum of the newborn pig and suggested that enhanced ileal LPH activity could be due to reduced cell migration rate with increased enterocyte life span.

To investigate the mechanism(s) by which insulin stimulates prematurely BBM enzymes in rat immature enterocytes, we have assessed the responses to insulin of four BBM hydrolases (SI, maltase, LPH, and aminopeptidase) and examined whether their increases in activity are dose-dependent and associated with corresponding dose-dependent increases in DNA synthesis rates, mRNA levels (SI and LPH), and intracellular polyamine concentrations. In addition, we have compared the effects on rat immature enterocytes of an analog of the hormone termed B-Asp10, which exhibits in vitro a 3.5-fold increase in binding affinity resulting in sustained signaling from the receptor and a 2-fold increase in metabolic potency over normal insulin(12, 13). Finally, we have examined the effects of IGF-I, given at doses equivalent to insulin and of anti-IR MAb on the activities of BBM enzymes that were found to be responsive to insulin.

METHODS

Materials. Human insulin was purchased from Novo Industries(Copenhagen, Denmark). Insulin B-Asp10, an analog of insulin in which histidine was replaced by aspartic acid in position 10 of the B-chain was a generous gift of Dr G. Danielsen (Novo Industries, Copenhagen, Denmark). Insulin B-Asp10 was produced by recombinant DNA techniques and by side-directed mutagenesis. Baker's yeast was used for the production and purification of the analog, as previously described(14, 15). IGF-I (human recombinant) was purchased from Boehringer Mannheim Biochemicals, (Mannheim, Germany). The peptide was found to be active in vitro on cells from humans, rats, mice, and chicken. Anti-rat IR MAbs were purchased from Amersham (Ghent, Belgium). The IR MAb recognizes epitopes of the α- and β-receptor subunits, binds competitively with insulin to the extramembranous domain of theα-receptor subunit(16), and inhibits in vitro autophosphorylation of the β-receptor subunit in immature rat intestinal membranes(9, 16).

Animals. Litters of Wistar rats, acclimatized to standard conditions of room temperature, light-dark cycles, and feeding schedules, were used throughout the study. To equalize conditions of nursing and lactation, control and experimental litters sizes were reduced either to six pups per mother or to eight pups per mother immediately after birth (d 0). During the nursing period, pups remained with their mother in individual cages and were weaned onto a pelleted diet (diet UAR 04, Villemoisson-sur-Orge, France) by d 15.

Treatment schedules. All studies were approved by the National and University Ethic Committees for animal welfare. By d 10 postpartum, pups were assigned at random to either control or experimental groups. Insulin or the analog B-Asp10 were injected intraperitoneally to sucklings, from d 11 until d 14 at the doses of 1 mU (0.04 μg), 4 mU (0.16 μg), or 12.5 mU(0.43 μg) per g of body weight/d, in a volume of 100 μL. These doses were divided into two injections (0730 and 1900 h) and were identical to those used in our previous studies(5, 9, 10). Control groups were treated following the same schedule with equivalent volumes of the vehicle.

To compare the effects of IGF-I with insulin, a dose of 0.1 μg of IGF-I per g of body weight/d was given by s.c. injections, twice daily from d 11 until d 14. This dose of IGF-I was chosen because s.c. administration of 1μg of IGF-I/d to suckling pups has been reported to enhance BBM enzymes but not intestinal growth(17). Control groups received either 0.1 μg (2.7 mU) of insulin per g of body weight/d divided into two s.c. injections or an equal volume of the vehicle. Anti-IR MAbs were administrated by i.p. injections, at the dose of 20 μg of protein per injection, given twice daily from d 11 until d 14 postpartum in a volume of 100 μL. A last dose was given on d 14, 2 h before sacrifice. Control groups received insulin (4 mU per g of body weight/d) or the vehicle according to the same schedule.

Preparation of tissues. On the day of the experiment (d 14), rats were killed by rapid decapitation, and the small intestine from the pylorus to the ileocecal valve was removed, trimmed of fat and mesentery, and rinsed with ice-cold saline. After being weighed and measured under fixed tension (0.1 g), the small intestine was divided into two equal segments. The proximal half was considered the jejunum and the distal half the ileum. Each segment was promptly opened lengthwise, and the mucosa was scraped off with glass slides. It was thereafter wrapped in Parafilm and frozen at -170°C until use. To collect epithelial cells from the crypt villus axis, we used Weiser's(18) method modified by Raul et al.(19). Briefly, a 10-cm jejunal segment was excised, everted, and rinsed in ice-cold PBS. By successive incubations (37°C, 10 min) in PBS containing 1.5 mM EDTA and 0.5 mM DTT (pH 7.4) and constant shaking at 170 rpm, 10 sequential fractions of cells were obtained. Villus cells were released first and were characterized by their high LPH activity, whereas crypt cells were identified by their low LPH activity and by their ability to incorporate [methyl-3H]thymidine. Villus cells(fractions 1-4) and crypt cells (fractions 5-10) were pooled, washed in PBS, and stored at -20°C in water until use.

Because changes in disaccharidase activities in response to insulin were found to be similar in the jejunum and ileum with as expected lower activities in the ileum BBM (P2 fraction) were isolated from the whole small intestinal mucosa and purified according to the method of Schmitz et al.(20). Preliminary assays revealed that LPH activity was 15 times higher in BBM than in the crude homogenate.

Biochemical determinations. Glycemia and plasma total corticosteronemia were measured by standard methods on trunk blood collected at the time of sacrifice, 2 h after the last injection of insulin. This timing was chosen because the hypoglycemic reaction and the potential stimulation of endogenous corticosterone release peaks 2 h after intraperitoneal injection of the hormone(9). SI, maltase, and LPH activities were assayed on isolated cell fractions and on purified BBMs by the method of Messer and Dahlqvist(21) within 1 wk of collection. Activities were expressed in micromoles of substrate hydrolyzed per min per g of protein. Aminopeptidase was determined by the method of Maroux et al.(22) using L-alanine-p-nitroanilide as a substrate. One unit equals 1 μmol of p-nitroaniline formed per min at 37°C. Specific activity of the enzyme was expressed in units per g of protein or in milliunits per mg of BBM proteins.

Protein content in cell homogenates and in BBMs was determined by the method of Lowry(23). The concentrations of putrescine, spermidine, and spermine in cell fractions were quantified by HPLC (Gilson model 306, mixing chambers, model 811A, Gilson) as previously described(24). Polyamine concentrations were established by comparing on a standard curve the integrated surface area of peaks with areas of dansylated polyamines of known concentration. Results were expressed as nanomoles per mg of cell protein.

DNA synthesis. Incorporation rates of[methyl-3H]thymidine into DNA were estimated by the administration (i.p.) of 2 μCi·g of body weight-1 of[methyl-3H]thymidine (specific activity, 5000 Ci/mmol, Amersham, Ghent, Belgium). To circumvent circadian periodicity in DNA synthesis rates(25), the tracer was given exactly 2 h before sacrifice at the same hour of the day (0700 h). DNA was quantified in mucosal homogenates by the method of Burton(26) modified by Giles and Myers(27) using calf thymus DNA as a standard. DNA concentrations were calculated per g of mucosa and per cm of gut length. Incorporation rates were expressed as count/min per mg of DNA and per cm of gut length.

cDNA and oligonucleotide probes. The plasmid PUC19 containing a 3-kb rat SI cDNA probe (G.C.4-5) was a gift of Dr. S. Henning (Departments of Pediatrics and Cell Biology, Baylor College of Medicine, Houston, TX)(28), whereas a 40-mer synthetic oligonucleotide probe for rat GAPDH was used to assess GAPDH mRNA levels as a control(29). The cDNA probes were isolated from vector DNA by appropriate restriction endonuclease digestion and agarose gel resolution, and were labeled by random priming using the multiprime DNA labeling system(Boehringer Mannheim) in the presence of [γ-32P]dATP for nucleotide probes or [α-32P]dCTP for cDNA probes.

RNA isolation and Northern hybridization analysis. Total cytoplasmic RNA was extracted from the mucosa scraped from a 10-cm segment located at the mid small intestine to minimize contamination by pancreatic RNases. The acid guanidium thiocyanate-phenol-chloroform method of extraction was used(30). RNA was fractionated by electrophoresis on 1% agarose, 3.7% formaldehyde gels (35 μg of RNA/lane). Blotting of RNA-agarose gel onto nylon membranes (Hybond-N, Amersham, Ghent, Belgium) was performed as described previously(31). The membranes were hybridized to specific 32P-labeled probes for GAPDH, SI(28), and LPH(31, 32) according to standard procedures. After washing, the membranes were exposed(96 h) to Kodak XAR-5 films, using an intensifying screen. Experiments were repeated twice.

Statistical methods. Differences between means were tested for statistical significance (p < 0.05) using ANOVA and where appropriate by a t test. If not appropriate, differences between means were tested by the nonparametric Mann-Whitney U test.

RESULTS

Suckling pups treated with hormones (insulin, B-Asp10, IGF-I) or with their respective vehicles (controls) appeared healthy without external signs of malnutrition. Growth data and mucosal mass parameters recorded in sucklings treated either with IGF-I (0.1 μg per g of body weight/d) or with equivalent doses of insulin (0.1 μg or 2.7 mU per g of body weight/d) are presented in Table 1. Compared with controls, final body weights, as well as intestinal length, intestinal weight, and mucosal weight, expressed per cm of gut length, were equivalent in IGF-I-, insulin-, and vehicle-treated groups. Regarding the responses of BBM hydrolases(Table 2), IGF-I failed to enhance the activity of the four BBM enzymes tested, whereas at the same dose insulin did increase significantly SI and maltase specific activity but had no effect on LPH and aminopeptidase. Although insulin-treated rats exhibited a transient hypoglycemia 2 h after the last injection of insulin, plasma total corticosteronemia remained low and equivalent in treated rats and controls, confirming our previous observations on adrenalectomized rats that the effects of insulin are not mediated by endogenous release of corticosterone(9, 10). Active levels of plasma corticosterone measured in weanling rats are at least 15 times higher than the levels measured here in sucklings(9). This was further confirmed by measuring the enzyme activities in isolated villus (Fig. 1) and crypt cells (Fig. 2) in response to low and high doses (1 and 12.5 mU) of insulin. Low doses of insulin (1 mU) were found to be devoid of hypoglycemic reaction(10). As shown in Figure 1, SI and maltase activities did increase in villus cells according to a dose-dependent response to the hormone. Likewise, the intracellular concentration of spermidine and of spermine did increase gradually according to the dose of insulin injected. There was, however, no effect of insulin on the concentration of putrescine. A similar pattern of enzyme response to insulin was observed in crypt cells including dose-related increases in SI and maltase activities. Changes in spermidine and in spermine concentrations were less marked in crypt cells than in villus cells and were not significant.

Table 1 Growth parameters in suckling rats (d 14) treated with insulin, IGF-I, or saline
Full size table
Table 2 Response of BBM hydrolases from rat immature enterocytes (d 14) to insulin, IGF-I, or saline
Full size table
Figure 1
figure 1

In villus cells, changes in SI and maltase-specific activities and in the concentrations of putrescine, spermine, and spermidine in response to different doses of insulin [1 (▪, n = 20] and 12.5 (□, n = 20) mU per g of body weight). Enzyme activities and polyamine concentrations were measured in villus cells isolated from the jejunum of 14-d-old suckling pups treated with insulin or with its vehicle[controls, □, n = 20]. For each statistical comparison, the probability level (p) is below the indicated value; data are mean± SE; NS = not significant.

Full size image
Figure 2
figure 2

In crypt cells, changes in SI and maltase specific activities and in the concentrations of putrescine, spermine, and spermidine in response to different doses of insulin [1 ▪, n = 19] and 12.5 [□, n= 18] mU per g of body weight). Enzyme activities and polyamine concentrations were measured in crypt cells isolated from the jejunum of 14-d-old sucklings pups treated with insulin or with its vehicle [controls, □, n = 16]. For each statistical comparison, the probability level (p) is below the indicated value; data are mean ± SE; ns = not significant.

Full size image

At the dose of 1 mU per g of body weight/d, insulin enhanced moderately but not significantly intestinal weight, mucosal mass, mucosal DNA, and incorporation rates of [methyl-3H]thymidine into DNA, measured in the jejunum and ileum (Table 3). After treatment with 12.5 mU per g of body weight, all these parameters exceeded markedly the values measured in controls. This demonstrates that exogenous administration of insulin can exert in vivo a mitogenic effect on the immature small intestine that appears to be dose-dependent like as SI and maltase but results in significant changes in gut mass only for high doses of the hormone (12.5 mU per g of body weight/d).

Table 3 Changes in mucosal mass parameters, mucosal DNA, and [methyl-3H]-thymidine incorporation in mucosal DNA of sucklings (d 14) treated with insulin (INS) at different doses
Full size table

The specific regulation of SI and maltase by the insulin receptor is shown in Table 4. Compared with vehicle-treated controls, both enzymes were significantly increased in BBMs from insulin-treated rats (4 mU per g of body weight/d) but they were decreased when rats were treated with IR MAb, indicating that the signal regulating their activity is dependent upon the binding of endogenous insulin to its own receptor. In concordance with the data presented in Table 2, administration of IR MAb or of insulin (4 mU per g of body weight/d) did not affect LPH activity. Also, the BBM protein concentration remained equivalent in the three groups of rats studied.

Table 4 Response of rat immature enterocytes (day 14) to anti-IR MAb
Full size table

To determine whether the duration and magnitude of the insulin receptor-mediated signal might be of importance for the biologic response, we compared the enzyme responses to insulin and to the analog B-Asp10. The results are shown in Figure 3. Compared with equivalent doses of insulin (12.5 mU) B-Asp10 caused an overexpression of SI activity by 3.5-fold and of maltase activity by 1.5-fold. There was, however, no difference between insulin and B-Asp10 regarding the stimulation of LPH and aminopeptidase, although the activity levels reached by these enzymes were significantly higher than those measured in control rats.

Figure 3
figure 3

Changes in SI, maltase, aminopeptidase, and LPH specific activity in response to insulin and to the analogue B-Asp10. Insulin and the analog B-Asp10 were administered from d 10 to 14 at the same dose (12.5 mU·g of body weight/d). Enzyme activities were measured in BBM samples purified from the whole intestinal mucosa of preweaning rats. Control rats were treated with an equivalent volume (100 μL) of vehicle alone; data are mean ± SE. □, controls (n = 8); ▪, 12.5 mU insulin (n = 8); and □, 12.5 mU Asp10, n = 8].

Full size image

Finally, we have assessed the effects of insulin on the expression of SI and LPH mRNAs. Before hybridization, total cytoplasmic RNA isolated from the jejunal mucosa of insulin-treated pups (4 and 12.5 mU·g of body weight/d) or of controls was fractionated by 1% agarose/3.7% formaldehyde gel electrophoresis and transferred onto Hybon-N membranes. Staining of membranes in methylene blue (40 mg/L in 0.5 M sodium acetate, pH 5) followed by washing revealed that for each deposit (35 μg of RNA/lane), 18 and 28 S ribosomal RNAs were found to be equivalent in abundance (not shown). The autoradiograms depicted in Figure 4 show that, in control rats, SI mRNA was not detected (lane 1), whereas it accumulated in insulin-treated rats (lanes 2 and 3) proportionally to the dose of insulin administered (4 and 12.5 mU·g of body weight/d). LPH mRNA levels were increased by 2-fold (densitometry) in insulin-treated rats (lanes 2 and 3) compared with controls (lane 1). The changes in LPH mRNA were not proportional to the changes in LPH activity, because at 4 mU, insulin did not increase the enzyme activity in BBMs (Table 4), whereas at 12.5 mU, the enzyme activity was significantly stimulated (Fig. 3) without a corresponding increase in LPH mRNA (Fig. 4). GAPDH was used as a positive control because stimulation of GAPDH gene transcriptional activity is known to be a specific effect of insulin(29). Using densitometric measurements, we found that the mean increases in GAPDH mRNA levels in insulin-treated rats were respectively of +44% (lane 2) and + 257% (lane 3) over controls (lane 1).

Figure 4
figure 4

Insulin-induced changes in mRNA expression of SI, LPH, and GAPDH. Total cytoplasmic RNA was extracted from the intestinal mucosa of 14-d-old rats (6/litter) treated with insulin at 4 mU (n = 6)(lanes 2) or at 12.5 mU (n = 6) (lanes 3) per g of body weight/d. Controls (n = 6) received the hormone vehicle according to the same schedule (lanes 1). Each lane was charged with 35 μg of total RNA. After transfer onto Hybon-N membranes, mRNAs were revealed by hybridization to specific 32P-labeled cDNA probes for rat SI and LPH, and to 32P-labeled oligonucleotide probe for GAPDH. Changes in LPH mRNA and GAPDH mRNA abundance were measured by densitometry in arbitrary units.

Full size image

DISCUSSION

Using different experimental approaches, the present study demonstrates that the precocious expression of SI and the stimulation of maltase in rat immature enterocytes in response to insulin is triggered by a dose-dependent signal mediated by the insulin receptor. First, low doses of exogenous insulin were able to enhance the activities of both enzymes, whereas at the same dose IGF-I had no detectable effect, suggesting that the signal is not mediated by the IGF-I receptor. Second, when pups were treated with IR MAbs that bind competitively in place of endogenous insulin to the α-receptor subunit(9, 16) SI and maltase activities were significantly reduced in BBMs compared with the activity levels of these enzymes measured in controls and in insulin-treated rats. Third, administration of the analogue B-Asp10, which displays in vitro a 3.5-fold increase in receptor binding affinity(12–14), resulted in an overexpression of SI (3.5-fold) and of maltase (1.5-fold) compared with equivalent doses of normal insulin. The main difference between insulin and B-Asp10 is a more intense and prolonged phosphorylation of the receptor tyrosine kinase and of certain intracellular substrates such as the 52-kD Shc(12). In accordance with our results, recent work(12) has shown that B-Asp10 enhances 3-O-methyl glucose uptake and lipogenesis in cultured Chinese hamster ovary cells overexpressing insulin receptors, by, respectively, 2.2- and 2.0-fold over normal insulin. This indicates that the duration and magnitude of the receptor-mediated insulin signal could be of critical importance for the selection and amplitude of the biologic response. An original finding of our study is that the pattern of responses of LPH and aminopeptidase activities to insulin, B-Ap10, and IR MAb differed substantially from that of SI, SI mRNA, and maltase. Several studies(31–33) have established that the transcriptional activity of SI and LPH genes are regulated differently in rat small intestine during growth and that developmental changes in SI are correlated with the amount of enzyme protein and its mRNA. The present findings are in concordance and show a parallelism between SI activity levels and SI mRNA accumulation in response to different doses of insulin. In contrast, this was not the case for LPH activity and LPH mRNA. In concordance with other studies(11, 34), we found no relationship between LPH activity and abundance of LPH mRNA. As suggested, translational and posttranslational mechanisms must be of major importance in developmental regulation of lactase(34) and in its response to insulin, a hormone that regulates by different molecular pathways both transcriptional and translational mechanisms. The specificity of insulin-induced gene transcription was attested by changes in GAPDH mRNA levels that control the synthesis of the glycolytic enzyme GAPDH(29). The changes documented here in GAPDH mRNA in response to insulin are of particular interest, because GAPDH mRNA is largely used as a stable internal control to quantify by densitometry the abundance of specific mRNAs. The selective responses of SI and maltase to low doses of insulin and their inhibition by IR MAb further confirms the physiologic role of insulin in maturation of the rat small intestine. Ontogenic increases in SI and maltase activities occur at a critical period of time (d 14-18) during which immature enterocytes exhibit marked responsiveness to insulin(5). At this time, plasma insulin levels rise actively(3) whereas milk-borne insulin is still present, allowing optimal interaction of the hormone with intestinal receptors whose concentration is higher than at any other time and which are located on both endoluminal and basolateral membranes of the cell. After weaning, the decrease by 2-fold in receptor concentration(35) would explain unresponsiveness of mature enterocytes to the hormone.

Although mucosal mass parameters and DNA synthesis rates were prematurely enhanced by insulin, induction of SI and stimulation of maltase by low doses of insulin cannot be explained by the mitogenic effect of insulin because after a single low dose (1 mU) of insulin, SI activity is detected very rapidly (within 6 h, corresponding to the time of half-life of the enzyme) in both villus and crypt cells, whatever their position on the villus crypt axis and their degree of differentiation(10). In agreement, Nsi-Emvo et al.(36) have reported that precocious expression of SI in rat immature intestine is unrelated to intestinal cell turnover. Alternately, induction of SI by activation of proto-oncogenes consecutive to starvation(36) can also be excluded because sucklings were not malnourished nor killed after a fasting period.

During the weaning period, the rise in intestinal ODC, and the increases(5-fold) in dietary polyamines and in intracellular polyamine concentration play an essential role in intestinal mucosal maturation(9, 37, 38). We have previously shown that oral administration of spermine or spermidine to suckling rats induces precocious enzymatic maturation of the small intestine, the effects being dose-dependent(38). Oral administration of spermine also induces the accumulation of SI mRNA in the immature small intestine(39). Furthermore, inhibition of ODC, the key enzyme of the de novo synthesis of polyamines, by administration to sucklings of α- difluoromethylornithine delays the onset of maturation(40). These observations indicate that changes in intracellular polyamines might mediate the effects of maturation-inducing effectors such as glucocorticoids(41) and epidermal growth factor. In this study, we present evidence that exogenous insulin produced significant increases in spermine and spermidine concentrations that appeared to be dose-dependent at least in isolated villus cells. In addition, the differences measured in villus cells of spermine (45%) and spermidine concentrations (55%) in response to insulin are consistent enough to directly influence gene transcription and enzyme expression, because they were even greater than the changes measured in mucosal spermine and spermidine after oral administration of these substances(42).

Although we cannot provide direct evidence that in response to insulin the intracellular increases in polyamines mediate the transcription of SI gene, there is a striking parallelism between the exogenous administration of insulin and spermine or spermidine to infant rats regarding the dose-dependent responses of SI and maltase to these stimuli(9, 38). The mechanism(s) by which insulin enhances the intracellular concentration of polyamines is not completely elucidated. Our preliminary findings(9) indicate that the insulin signal is not mediated by an activation of ODC, because after complete inhibition of ODC and ODC mRNA by α- difluoromethylornithine, immature enterocytes remained sensitive to insulin. In contrast, we found that insulin stimulated markedly the endoluminal uptake of [14C]spermine in BBM vesicles, which in turn results in accumulation of these substances in the cell. Similar observations have been reported with epidermal growth factor which stimulates in vitro the uptake of polyamines in cultured CACO-2 cells by the possible translocation of a specific carrier from the cytoplasm into the cell membrane(43).

Thus the rise in insulin acts as a primary signal transduced into the cell by binding of the hormone to its intestinal receptor followed by phosphorylation of the receptor tyrosine kinase, which indirectly triggers the premature expression of SI mRNA through a downstream cascade of intracellular signaling molecules. The precise molecular mechanism by which this signal affects the uptake of endoluminal polyamines and their intracellular concentration remains unknown. Although intracellular insulin can stimulate RNA and protein synthesis(44) and accumulate into nuclei(45), the target of insulin can hardly be the SI gene, because the region of rodent gene promoter whose sequence has been determined(30, 46) does not show any characteristic binding site for insulin nor for its receptor. Recently, Traber et al.(46) proposed a model of the complex SI regulatory elements in which transcription of SI gene might be under the control of intestinal specific nuclear proteins binding to the SIF1 site which is a specific nucleotide sequence located just upstream of the transcription start site of the SI gene. A recent work(34) has confirmed that SIF1 binding proteins of different molecular weights regulate SI but not LPH expression during postnatal development. There are likely many other DNA regulatory proteins and additional elements located outside the SI promoter that are important for directing developmental SI gene transcription. In our model of premature induction of SI expression, the insulin receptor-mediated signal might represent, outside the promoter, one specific event leading to activation of the SI gene. A possible mechanism would be that polyamines regulate the expression of genes implied in the synthesis of nuclear proteins binding to regulatory elements directing SI gene transcription.