Establishment of polarized endocytosis in differentiable intestinal HT29-18 subclones

Subclones of the HT29-18 clone, derived from a human adenocarcinoma, are able to acquire an enterocyte-like phenotype depending on the culture conditions. To investigate fluid-phase and receptor-mediated endocytosis in the polarized subclone HT29-18-C1, we established culture conditions that allowed cell growth on permeable supports. HT29-18-C1 monolayers had an electrical resistance of 43 ohms.cm2 and developed a transepithelial potential of about 2 mV. Transferrin receptors were uniformly distributed on the entire cell surface of undifferentiated HT29-18 cells but were located on the basolateral membrane of differentiated cells. Transferrin had a high affinity (Kd = 2.5 x 10(-9) M) for its receptor independent of the state of differentiation. The number of transferrin receptors and the mRNA amounts encoding them were comparable in the undifferentiated and differentiated HT29-18 cells. Transferrin was quickly internalized and recycled back to the cell surface of undifferentiated HT29-18 cells. The same phenomenon also occurred in differentiated HT29-18 cells, but the receptors were limited to the basolateral membrane. In the presence of ammonium chloride, the process was slower but remained polarized. Fluid-phase uptake was also investigated with horseradish peroxidase (HRP) in differentiated HT29-18 C1 cells. HRP that was internalized in 1 hour from a given membrane domain preferentially recycled back to the same membrane domain. No significant accumulation of the enzyme in the late endosomes and lysosomes of the differentiated HT29-18-C1 cells was observed.     

Human platelets are defective in processing of cholera toxin.

Cholera toxin is unable to elevate cyclic AMP levels in intact human platelets despite being very efficacious in this respect in other mammalian cells; in the presence of 0.5 mM-isobutylmethylxanthine, we found that 3-6nM-cholera toxin over 3h at 37 degrees C elevated platelet cyclic AMP from 33 +/- 13 to 39 +/- 12pmol/mg of protein (means +/- S.D.; n = 12). We have investigated the basis for this lack of response. 125I-labelled cholera toxin bound to platelets both saturably and with high affinity (Kd congruent to 60pM; Bmax. congruent to 50fmol/mg of protein). Incubation of platelets with the putative cholera toxin receptor monosialoganglioside GM1 enhanced 125I-labelled cholera toxin binding at least 40-fold but facilitated only a minimal (less than or equal to 3-fold) elevation of platelet cyclic AMP levels. In contrast, dithiothreitol-activated cholera toxin markedly stimulated adenylate cyclase activity in platelet membranes. Platelet cytosol both enhanced stimulation of adenylate cyclase activity by activated cholera toxin (A1 subunit) and supported stimulation by the A1-A2 subunit of cholera toxin. Neither GTP nor NAD+, both necessary for response to cholera toxin, was lacking in intact platelets. However, we found that platelets were unable to cleave cholera toxin to the active A1 subunit (as assessed by sodium dodecyl sulphate/polyacrylamide-gel electrophoresis). By contrast, murine S49 lymphoma cells were able to generate the A1 subunit with a time course that closely resembled the kinetics of toxin-mediated cyclic AMP accumulation in these cells. Thus we conclude that human platelets are defective in their ability to process surface-bound cholera toxin. These results indicate that binding of cholera toxin to surface receptors is necessary, but not sufficient, for expression of the toxin effect and the generation of the A1 subunit of the toxin may be rate-limiting for expression of cholera toxin response.     

Induction of refractoriness to isoproterenol by prior treatment of C6-2B rat astrocytoma cells with cholera toxin.

Rat C6-2B astrocytoma cells responded to cholera toxin treatment with an 8-fold increase in intracellular cyclic AMP concentrations. Cyclic AMP levels began to rise 60--90 minutes after addition of the toxin and reached maximal concentrations in 3 hours. Cells exposed to cholera toxin and the phosphodiesterase inhibitor, 1-methyl-3-isobutylxanthine (MIX), displayed an increase in cyclic AMP of 15-fold. The peak isoproterenol response was reduced 80--90% in cells previously treated with cholera toxin. Cholera toxin-induced refractoriness was time dependent and was not altered by concurrent treatment with propranolol. Prolonged exposure of the cells to isoproterenol reduced the cyclic AMP response to cholera toxin by 80%. MIX augmented both cholera toxin-induced refractoriness and isoproterenol-induced refractoriness. Cycloheximide inhibited the full development of refractoriness to both cholera toxin and isoproterenol. These results indicate that C6-2B cell refractoriness to cholera toxin is mediated by cyclic AMP and requires new protein synthesis. Refractoriness in C6-2B cells does not appear to be agonist-specific and probably involves a common locus of action on adenylate cyclase beyond that of the membrane receptors for cholera toxin and isoproterenol.     

Action of cholera toxin on dispersed acini from guinea pig pancreas.

In dispersed acini from guinea pig pancreas cholera toxin bound reversibly to specific membrane binding sites to increase cellular cyclic AMP and amylase secretion. Cholera toxin did not alter outflux of 45Ca or cellular cyclic AMP. Binding of 125I-labeled cholera toxin could be detected within 5 min; however, cholera toxin did not increase cyclic AMP or amylase release until after 40 min of incubation. There was a close correlation between the dose vs. response curve for inhibition of binding of 125I-labeled cholera toxin by native toxin and the action of native toxin on cellular cyclic AMP. With different concentrations of cholera toxin, maximal stimulation of amylase release occurred when the increase in cellular cyclic AMP was approximately 35% of maximal. Cholera toxin did not alter the increase in 45Ca outflux or cellular cyclic GMP caused by cholecystokinin or carbachol but significantly augmented the increase in cellular cyclic AMP caused by secretin or vasoactive intestinal peptide. The increase in amylase secretion caused by cholera toxin plus secretin or vasoactive intestinal peptide was the same as that with cholera toxin alone. On the other hand, the increase in amylase secretion caused by cholera toxin plus cholecystokinin or carbachol was significantly greater than the sum of the increases caused by each agent alone.     

Cytokine-induced upregulation of NF-kappaB, IL-8, and ICAM-1 is dependent on colonic cell polarity: implication for PKCdelta

As described for a long time, carcinoma-derived Caco-2 cells form a polarized epithelium in culture, whereas HT29-D4 cells are nonpolarized and undifferentiated but can form a polarized monolayer when cultured in a galactose-supplemented medium. Using NF-kappaB translocation and IL-8 and ICAM-1 gene activation as an index, we have studied the relationship between the differentiation state and the cell response to cytokines. We found that differentiated Caco-2 and HT29-D4 cells were responsive to both cytokines TNFalpha- and IL-1beta-mediated activation of NF-kappaB but that undifferentiated HT29-D4 cells were unresponsive to IL-1beta. However, the expression of endogenous ICAM-1 and IL-8 genes was upregulated by these cytokines in either cell lines differentiated or not. Upregulation of ICAM-1 gene occurred when IL-1beta or TNFalpha was added to the basal, but not apical surface of the differentiated epithelia. Finally, it appeared that in polarized HT29-D4 cells, the IL-1beta-induced translocation of NF-kappaB was connected to PKCdelta translocation.     

Action of cholera toxin on dispersed acini from guinea pig pancreas

In dispersed acini from guinea pig pancreas cholera toxin bound reversibly to specific membrane binding sites to increase cellular cyclic AMP and amylase secretion. Cholera toxin did not alter outflux of ⁴⁵Ca or cellular cyclic AMP. Binding of ¹²⁵I-labeled cholera toxin could be detected within 5 min; however, cholera toxin did not increase cyclic AMP or amylase release until after 40 min of incubation. There was a close correlation between the dose vs. response curve for inhibition of bindind of ¹²⁵I-labeled cholera toxin by native toxin and the action of native toxin on cellular cyclic AMP. With different concentrations of cholera toxin, maximal stimulation of amylase release occurred when the increase in cellular cyclic AMP was approximately 35% of maximal. Cholera toxin did not alter the increase in ⁴⁵Ca outflux or cellular cyclic GMP caused by cholecystokinin or carbachol but significantly augmented the increase in cellular cyclic AMP caused by secretion or vasoactive intestinal peptide. The increase in amylase secretion caused by cholera toxin plus secretin or vasoactive intestinal peptide was the same as that with cholera toxin alone. On the other hand, the increase in amylase secretion caused by cholera toxin plus cholecystokinin or carbachol was significantly greater than the sum of the increases caused by each agent alone.     

Differentiation of a clone isolated from the HT29 cell line: polarized distribution of histocompatibility antigens (HLA) and of transferrin receptors

The HT29 cell line, derived from a human colon adenocarcinoma, is able to differentiate if galactose replaces glucose in the culture medium. We have isolated a clone (HT29-18) from this cell line which displays differentiated properties of the parent cell line. HT29-18 cells grown in glucose-containing medium form multiple layers of round cells without specific cell-cell adhesion. In contrast, when grown in galactose-containing medium, they form a monolayer with tight junctions and exhibit a well differentiated brush border at their apical membrane, which faces the culture medium. The polarized properties of HT29-18 cells grown in galactose-containing medium were demonstrated by immunofluorescent techniques with antibodies against 2 plasma membrane proteins. Class I histocompatibility antigens (HLA) and transferrin receptors, 2 well characterized integral membrane proteins, are uniformly distributed on the cell surface of undifferentiated HT29-18 cells, but acquire a polarized distribution during differentiation, localized on the basolateral membranes and absent from the apical surface. Binding of 125I-labeled transferrin was used to determine transferrin receptor distribution on apical and basolateral membranes. Functional tight junctions in the differentiated cultures were demonstrated, as the monolayer was impermeable to a permeation dye (ruthenium red) as well as to antibodies. The sealing of these tight junctions is, as in vivo, Ca++-dependent as they could be opened by a short incubation in Ca++-free medium.     

Restricted localization of functional vasoactive intestinal peptide (VIP) receptors in in vitro differentiated human colonic adenocarcinoma cells (HT29-D4)

HT29-D4, a clone of the human colonic adenocarcinoma cell line (HT29), possesses at its cell surface specific binding sites for the vasoactive intestinal peptide (VIP) (KD = 0.5 nM). Their molecular weight was previously estimated to 117 kDa and 64 kDa. This clone underwent functional and structural differentiation when grown in a glucose-free galactose-containing medium. The [125I]VIP binding capacity of cells grown in this medium gradually declined while the cell density increased and reached a value close to zero when cell monolayer was able to form hemicysts. At this time, cells presented numerous tight junctions and desmosomes and a well organized brush border. Binding capacity could be recovered when the post-confluent monolayers were previously disaggregated with EDTA. Neither the affinity for VIP nor the molecular weight of the [125I]VIP cross-linked polypeptides were modified in these cells compared to cells grown in glucose-containing medium. However, surface receptor number of differentiated cells was twice that of undifferentiated cells. Leakproof differentiated cell monolayers grown on permeable substratum produced cAMP in response to VIP only when the peptide was present in the lower chamber of the culture wells. Taking these data altogether, we conclude that the localization of functional VIP receptors is restricted to the basolateral domain in differentiated post-confluent HT29-D4 cells.     

Inhibition of cholera toxin-stimulated intestinal epithelial cell adenylate cyclase by adenosine analogs.

The ability of various adenosine analogs to inhibit cholera toxin activation of the intestinal epithelial cell adenylate cyclase-cyclic AMP system was investigated. After incubation of cells with cholera toxin for 6 hr, large increases in cellular cyclic AMP content were observed. Addition of 2', 5'-dideoxyadenosine during the last 30 min of this 6-hr incubation resulted in 70% reduction in elevated cyclic AMP content. Other analogs were not effective inhibitors. 2', 5'-Dideoxyadenosine was also a potent inhibitor of cholera toxin-activated intestinal cell adenylate cyclase activity with half-maximal inhibition occuring at 16 muM. NaF-stimulated cyclase was less susceptible to inhibition. The data suggest that inhibition by 2', 5'-dideoxyadenosine is due at least in part to direct inhibition of the cholera toxin-activated intestinal adenylate cyclase activity.     

Absorptive and mucus-secreting subclones isolated from a multipotent intestinal cell line (HT-29) provide new models for cell polarity and terminal differentiation

A clone HT29-18 has been isolated from the parent cell line HT-29, which derived from a human colon adenocarcinoma (Fogh, J., and G. Trempe, 1975, Human Tumor Cells in Vitro, J. Fogh, editor, Plenum Publishing Corp., New York, 115-141). This clone is able to differentiate as the parent cell line does. Differentiation occurs when glucose is replaced by galactose in the culture medium (Pinto, M., M.D. Appay, P. Simon-Assman, G. Chevalier, N. Dracopoli, J. Fogh, and A. Zweibaum, 1982, Biol. Cell., 44:193-196). We demonstrate here that the differentiated cloned population HT29-18/gal is heterogenous: although 90% of the cells show morphological characteristics of "absorptive cells", only 20-30% of them display sucrase-isomaltase in their apical microvillar membranes. About 10% of the entire cell population consists of cells containing mucous granules similar to intestinal goblet cells. We have isolated two subclones, HT29-18-C1 and HT29-18-N2, from the differentiated HT29-18/gal cells. HT29-18-C1 cells show morphological characteristics of polarized absorptive cells, when growing either in glucose- or in galactose-containing media, but the sucrase-isomaltase is not expressed in the cells grown in glucose-containing medium. The clone HT29-18-N2 is also polarized in both culture conditions and is similar to globlet cells in vivo. It grows as a monolayer, exhibits tight junctions, and contains numerous mucous granules whose exocytosis can be triggered by carbachol, a parasympathomimetic drug. We conclude that the clone HT29-18 first isolated was a multipotent cell population from which we isolated several subclones that differentiate either as absorptive (HT29-18-C1) or as mucous (HT29-18-N2) cells. In contrast to the parent HT-29 cell line, the subclones retain most of their differentiated properties in glucose-containing medium.     

Potentiation of cholera toxin-stimulated cyclic AMP production in cultured cells by inhibitors of RNA and protein synthesis

Cyclic AMP increased 8- to 10-fold after a 3-h treatment with 6 nM cholera toxin in rat C6-2B astrocytoma cells. In the presence of cycloheximide, cholera toxin increased intracellular cyclic AMP about 50-fold. Qualitatively similar potentiation of cholera toxin action by cycloheximide was observed in isolated swine aortic vascular smooth muscle cells. Cycloheximide, by itself, had no effect upon cyclic AMP levels and did not alter the apparent Ka for cyclic AMP generation by cholera toxin in the cells. Also, cycloheximide did not appear to augment cholera toxin action via inhibition of cyclic nucleotide phosphodiesterase. Puromycin and actinomycin D also augmented cholera toxin action in C6-2B cells. Potentiation of cholera toxin-increased cyclic AMP formation by cycloheximide was correlated with the inhibition of [14C]leucine incorporation into protein. These results indicate that the ability of cholera toxin to stimulate cyclic AMP production in C6-2B astrocytoma and swine vascular smooth muscle cells is enhanced by inhibition of de novo protein synthesis.     

Regulation of GH3 pituitary tumour-cell adenylate cyclase activity by activators of protein kinase C.

The influence of protein kinase C (PKC) activation on cyclic AMP production in GH3 cells has been studied. The stimulation of cyclic AMP accumulation induced by forskolin and cholera toxin was potentiated by 4 beta-phorbol 12,13-dibutyrate (PDBu). Moreover, PDBu, which causes attenuation of the maximal response to vasoactive intestinal polypeptide (VIP), also induced a small right shift in the dose-response curve for VIP-induced cyclic AMP accumulation. PDBu-stimulated cyclic AMP accumulation was unaffected by pretreatment of cells with pertussis toxin or the inhibitory muscarinic agonist, oxotremorine. PDBu stimulation of adenylate cyclase activity required the presence of a cytosolic factor which appeared to translocate to the plasma membrane in response to the phorbol ester. The diacylglycerol-generating agents thyroliberin, bombesin and bacterial phospholipase C each stimulated cyclic AMP accumulation, but, unlike PDBu, did not attenuate the stimulation induced by VIP. These results suggest that PKC affects at least two components of the adenylate cyclase complex. Stimulation of cyclic AMP accumulation is probably due to modification of the catalytic subunit, whereas attenuation of VIP-stimulated cyclic AMP accumulation appears to be due to the phosphorylation of a different site, which may be the VIP receptor.     

Inhibition of lipolysis and cyclic AMP accumulation by adenosine analogues in hamster epididymal adipocytes exposed to cholera toxin

The effects of adenosine, N6-phenylisopropyl adenosine and 2',5'-dideoxyadenosine on lipolysis and cyclic AMP accumulation, in hamster adipocytes treated with cholera toxin, were studied. Cholera toxin caused an increase in lipolysis and cyclic AMP accumulation that was dependent upon the concentration of toxin and the length of time cells were exposed to the toxin. When N6-phenylisopropyl adenosine or 2',5'-dideoxyadenosine were present, the lipolytic and cyclic AMP responses to cholera toxin were inhibited. The adenosine analogues were equally effective inhibitors of lipolysis and cyclic AMP accumulation, when they were added 1 or 2 h after exposure to the toxin. Enzymatic removal of endogenously produced adenosine with adenosine deaminase potentiated both the lipolytic and cyclic AMP responses to cholera toxin. In addition, the inhibitory effects of N6-phenylisopropyl adenosine, 2'5'-dideoxyadenosine and clonidine on lipolysis and cyclic AMP were enhanced consequent to enzymatic removal of adenosine. These data show responses of intact fat cells to N6-phenylisopropyl adenosine, 2',5'-dideoxyadenosine or removal of endogenous adenosine and provide evidence for an adenosine sensitivity of fat cells exposed to cholera toxin.     

The B subunit of cholera toxin enhances DNA synthesis in rat hepatocytes induced by insulin and epidermal growth factor.

The B subunit of cholera toxin, which binds to ganglioside GM1, enhanced DNA synthesis in rat hepatocytes in primary culture induced by insulin and/or epidermal growth factor. The effect was dose-dependent, and whole cholera toxin, activating adenylate cyclase, showed a higher effect than the B subunit alone. The B subunit acted additively with other agents that also increase cyclic AMP levels. A competitive antagonist of cyclic AMP could not suppress the effect of the B subunit completely. These data suggest that the effect is independent of the cyclic AMP signal pathway, and that GM1 plays a role in hepatocyte proliferation.     

Down-regulation of gonadotropin and beta-adrenergic receptors by hormones and cyclic AMP

Loss of gonadotropin receptors in murine Leydig tumor cells and of beta-adrenergic receptors in rat glioma C6 cells occurred following exposure of the cells to human chorionic gonadotropin and isoproterenol, respectively. Down-regulation of receptors was mimicked in part by other agents that elevated cyclic AMP levels in the cells such as cholera toxin and dibutyryl cyclic AMP. Whereas agonist-mediated receptor loss was rapid and almost total, down-regulation by cyclic AMP was slower and less extensive. Down-regulation of receptors did not appear to be accompanied by loss of the regulatory and catalytic components of adenylate cyclase. Hormone-mediated down-regulation was preceded by desensitization of hormone-stimulated adenylate cyclase. In contrast, there was no evidence that cyclic AMP caused desensitization. Finally, loss of receptors induced either by agonists or cyclic AMP required protein synthesis as cycloheximide inhibited down-regulation. We conclude that down-regulation of receptors in these cells is a complex process involving both cyclic AMP-independent and -dependent events.     

On the mechanism of action of cholera toxin on isolated rat adrenocortical cells Comparison with the effects of adrenocorticotropin on steroidogenesis and cyclic AMP output

The effects of cholera toxin on isolated rat adrenocortical cells have been investigated. Both steroid and cyclic AMP output from adrenal cells were increased by the toxin in a dose dependent fashion. The concentration of toxin for half maximal stimulation for both of these responses was about 40 ng/ml. Maximal steroidogenesis and cyclic AMP output was obtained with similar concentrations of the toxin. A correlation was observed between the low amounts of cyclic AMP produced in response to all doses of cholera toxin and to physiologically significant concentrations of adrenocorticotropin (ACTH) (< 0.1 munit/ml; i.e. submaximal for steroidogenesis in this system). This was in direct contrast to the much higher levels of cyclic AMP generated by concentrations of ACTH greater than 1 munits/ml. Time course studies demonstrated a time-lag between toxin addition and steroid response of at least 40 min. Binding of cholera toxin to adrenal cells was rapid and was 90% complete within 15 min at both 37 and 0°C. These data indicate that most of the delay in response to cholera toxin is due to processes subsequent to the initial binding interaction. Following the initial delay the subsequent maximal rate of steroidogenesis brought about by cholera toxin was very similar to that obtained with a concentration of ACTH that was maximal for steroidogenesis. Significant increases in cyclic AMP levels were detected about 20 min before increased steroidogenesis was apparent. Possible explanations for this result are considered. The results presented indicate great potential use for cholera toxin in the study of adrenal steroidogenic control mechanisms, particularly at the level of receptor mechanisms and the role of cyclic AMP.     

Cyclic AMP potentiates down regulation of epidermal growth factor receptors by platelet-derived growth factor

Pretreatment of $\text{Balb}\text{c}\text{-3T3}$ cells with platelet-derived growth factor (PDGF) has been shown to decrease binding sites for ¹²⁵I-labelled epidermal growth factor (EGF) (1,2,3). Agents which elevate cellular cyclic AMP concentrations enhance this ability, and the change in EGF binding is inversely proportional to the elevation of cyclic AMP. In quiescent density arrested cells, the sensitivity of cells to down regulation of EGF receptors by PDGF is proportional to the cyclic AMP content of the cultures in three different cell lines. Agents which elevate cyclic AMP and which potentiate PDGF mediated heterologous down regulation of EGF receptors are able, like cholera toxin (3), to stimulate cells to synthesize DNA in defined medium in the absence of EGF. Down regulation of EGF receptors by PDGF in combination with agents elevating cyclic AMP effectively mimics the action of EGF.     

Characterization by Western Blotting of Mouse Intestinal Glycoproteins Bound by Escherichia coli Heat-Labile Enterotoxin Type I

Escherichia coli heat-labile enterotoxin type I (LT-I)-binding galactoproteins, which were not recognized by cholera toxin, were detected in intestinal epithelial cells of BALB/c mouse by Western blotting. Inhibitory studies using lectins and modifications of sugar chain suggest that LT-I recognizes certain mucin-type sugar chains containing the terminal Galβ1-3GalNAc sugar sequence in the galactoproteins. The terminal sugar sequence is identical to that of GM1 ganglioside, the well-documented functional receptor for LT-I.