Solubilization and hydrodynamic characterization of guanine nucleotide sensitive vasoactive intestinal peptide-receptor complexes from rat intestine

The purpose of this work was to solubilize vasoactive intestinal peptide (VIP) receptors from rat small intestinal plasma membranes and to analyze the nature and function of its molecular form(s) in a nondenaturing environment. Membranes were incubated with 3 nM 125I-VIP, washed, and treated with 1% Triton X-100. Chromatography on Sephadex G-50 showed that 60% of the extractable radioactivity was eluted with macromolecular components in the void volume. This radioactive material was dramatically reduced when 1 microM unlabeled VIP was present in the incubation medium or when membranes were pretreated with trypsin or dithiothreitol. Macromolecular components that had bound 125I-VIP were further chromatographed on Sephacryl S-300. Two peaks were observed: a major one (80%) and a minor one (20%) with Stokes radii of 5.2 and 3.1 nm, respectively. The labeling of both components was inhibited by unlabeled VIP or peptide with NH2-terminal histidine and COOH-terminal isoleucine amide (a VIP agonist). The presence of GTP (0.1 mM) in the incubation medium of membranes completely abolished the labeling of the 5.2-nm component but did not affect that of the 3.1-nm one. Moreover, GTP induced dissociation of 125I-VIP from the 5.2-nm component isolated by Sephacryl S-300 chromatography. This effect was time dependent and nucleotide specific. In contrast, GTP did not affect the stability of the 3.1-nm component. After cholera toxin catalyzed [32P]ADP-ribosylation of membranes, chromatography of solubilized material on Sephacryl S-300 showed that a peak of 32P radioactivity was coeluted with the 5.2-nm component.(ABSTRACT TRUNCATED AT 250 WORDS)     

Solubilization of active and stable receptors for vasoactive intestinal peptide from rat liver

Vasoactive intestinal peptide (VIP) receptors were solubilized from rat liver using the zwitterionic detergent CHAPS. Optimal conditions of solubilization were obtained with 5 mM CHAPS and 2.5 mg protein/ml. The binding of 125I-VIP to CHAPS extracts was time- and pH-dependent, saturable and reversible. The following order of potency of unlabeled VIP-related peptides for inhibiting 125I-VIP binding was observed: VIP greater than helodermin greater than peptide histidine isoleucine amide (PHI) greater than rat growth hormone releasing factor (rGRF) greater than secretin. This peptide specificity is identical to that of rat liver membrane-bound receptors. VIP binding activity in the CHAPS extract was destroyed by trypsin or dithiothreitol in accordance with the known sensitivity of membrane-bound receptors to these agents. VIP receptors in CHAPS extracts were stable for at least 5 days at 4 degrees C. Scatchard analysis of equilibrium binding data indicated the presence in CHAPS extracts of high (H) and low (L) affinity binding sites with the following characteristics: KdH = 0.27 nM and BmH = 34 fmol/mg protein; KdL = 51 nM and BmL = 1078 fmol/mg protein. The guanine nucleotide GTP inhibited 125I-VIP binding to soluble receptors and enhanced the dissociation of soluble VIP-receptor complexes, suggesting that GTP-binding proteins were functionally associated with VIP receptors in solution. Gel filtration of solubilized VIP receptors on Sephacryl S-300 revealed a single binding component with a Stokes radius of 6.1 nm. It is concluded that active VIP receptors can be extracted from liver membranes by CHAPS. The availability of this CHAPS-soluble, stable and functional receptor from a tissue which can be obtained in large amounts represents a major step toward the purification of VIP receptors.     

Autoradiographic visualization of CNS receptors for vasoactive intestinal peptide

Receptors for VIP were characterized in the rat CNS. 125I-VIP bound with high affinity to rat brain slices. Binding was time dependent and specific. Pharmacology studies indicated that specific 125I-VIP binding was inhibited with high affinity by VIP and low affinity by secretin and PHI. Using in vitro autoradiographic techniques high grain densities were present in the dentate gyrus, pineal gland, supraoptic and suprachiasmatic nuclei, superficial gray layer of the superior colliculus and the area postrema. Moderate grain densities were present in the olfactory bulb and tubercle, cerebral cortex, nucleus accumbens, caudate putamen, interstitial nucleus of the stria terminalis, paraventricular thalamic nucleus, medial amygdaloid nucleus, subiculum and the medial geniculate nucleus. Grains were absent in the corpus callosum and controls treated with 1 microM unlabeled VIP. The discrete regional distribution of VIP receptors suggest that it may function as an important modulator of neural activity in the CNS.     

Characterization of receptors for vasoactive intestinal peptide solubilized from the lung

The zwitterionic detergent CHAPS was used to solubilize functional receptors for vasoactive intestinal peptide (VIP) from guinea pig lung. The solubilized receptors were resolved by high performance gel filtration in 3 mM CHAPS into two active fractions with apparent Stokes radii of 5.9 +/- 0.1 and 2.3 +/- 0.1 nm. The binding of 125I-VIP to the two receptor fractions was time-dependent, reversible, and saturable. Trypsin destroyed the binding activity of the receptor fractions, indicating their proteinic nature. Unlabeled VIP competitively displaced the binding of 125I-VIP to the 5.9-nm fraction (IC50 = 240 pM) and the 2.3-nm fraction (IC50 = 1.2 microM). Scatchard analysis indicated a single class of binding sites in each receptor fraction, with Kd values 300 pM and 0.97 microM for the 5.9- and 2.3-nm Stokes radii fractions, respectively. When the high affinity, 5.9-nm Stokes radius fraction was rechromatographed in 9 nM CHAPS, 46% of the binding activity eluted in the low affinity, 2.3-nm Stokes radius fraction, indicating that the latter is a product of dissociation of the high affinity receptor complex. GTP inhibited the binding of 125I-VIP to the high affinity complex but not the low affinity species. Scatchard plots of VIP binding by the high affinity receptors treated with GTP suggested the presence of two distinct binding sites (Kd 4.4 and 153 nM), compared to a single binding site (Kd = 0.3 nM) obtained in untreated receptors. The nonhydrolyzable GTP analog, guanyl-5'-yl-imidodiphosphate, inhibited VIP binding by the high affinity receptor fraction with potency nearly equivalent to that of GTP. These observations suggest that GTP-binding regulatory proteins are functionally coupled to the VIP-binding subunit in the high affinity receptor complex. The peptide specificity characteristics of the two receptor fractions were different. Peptide histidine isoleucine and growth hormone releasing factor, peptides homologous to VIP, were 87.5- and 22.9-fold less potent than VIP in displacing 125I-VIP binding by the high affinity receptor complex, respectively. On the other hand, growth hormone-releasing factor was more potent (22.7-fold) and peptide histidine isoleucine was less potent (31.3-fold) than VIP in displacing the binding by the low affinity species.     

Characterization of receptors for thyrotropin-releasing hormone-potentiating peptide on rat anterior pituitary membranes.

Previous studies (Bulant, M., Delfour, A., Vaudry, H., and Nicolas, P. (1988) J. Biol. Chem. 263, 17189-17196; Bulant, M., Roussel, J. P., Astier, H., Nicolas, P., and Vaudry, H. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 4439-4443) have shown that post-translational processing of rat thyrotropin-releasing hormone prohormone (pro-TRH) generates, besides thyrotropin-releasing hormone (TRH), a connecting decapeptide corresponding to prepro-TRH-(160-169), i.e. Ser-Phe-Pro-Trp-Met-Glu-Ser-Asp-Val-Thr. This peptide, which is named TRH-potentiating peptide (Ps4), is co-localized with TRH in the median eminence nerve endings and is involved in potentiation of the action of TRH on thyrotropin hormone release by pituitary in vitro and in vivo. To characterize the receptor(s) for TRH-potentiating peptide in the pituitary, a highly potent and metabolically stable derivative of Ps4, [I-Tyr0]Ps4, was radioiodinated. Binding of [125I-Tyr-0]Ps4 to rat pituitary membrane homogenates was specific, saturable, reversible, and linear with membrane protein concentration. Equilibrium measurements performed over a large range of concentrations revealed a single homogeneous population of high affinity binding sites (Kd = 0.22 nM; Bmax = 517 fmol/mg of membrane proteins). Several naturally occurring neuropeptides and hormones, including TRH, did not compete with [125I-Tyr0]Ps4 in the binding, which suggests the binding sites are specific to Ps4. Using C-terminal deletion analogs of [Tyr0]Ps4, we further showed the critical role the C-terminal residues Thr10, Val9, and Asp8 play in conferring high binding affinity and selectivity. Binding site tissue distribution and cross-reactivity binding studies suggest that the action of TRH-potentiating peptide is mediated through interaction with a specific pituitary cell-surface receptor which differ from those for TRH. [I-Tyr0]Ps4 reported in this paper, through its high binding affinity and specificity, its very low nonspecific binding, its high resistance to enzymatic degradation, and its high potentiating action in vitro should allow further progress in understanding the in vivo physiological function of Ps4.     

Functional receptors for vasoactive intestinal peptide in cultured vascular smooth muscle cells from rat aorta

Specific binding sites for vasoactive intestinal peptide (VIP), a potent vasodilatory polypeptide, and its effect on formation of intracellular cyclic AMP levels were studied in cultured vascular smooth muscle cells (VSMC) from rat aorta. Specific binding of ¹²⁵I-labeled-VIP to cultured VSMCs was time- and temperature-dependent. Scatchard analysis of binding studies suggested the presence of two classes of high and low affinity binding sites for VIP; the apparent K_d and the number of maximal binding capacity were ∼8×10⁻⁹ M and 60,000 sites/cell (high-affinity sites) and ∼4×10⁻⁸ M and 140,000 sites/cells (low-affinity sites), respectively. Unlabeled VIP competitively inhibited the binding of ¹²⁵I-labeled-VIP to its binding sites, whereas neither peptides structurally related to VIP, nor other vasoactive substances affected the binding. VIP stimulated formation of intracellular cyclic AMP in cultured VSMCs in a dose-dependent manner; the stimulatory effect of VIP on cyclic AMP formation was not blocked by propranolol and was additive with isoproterenol. The present study first demonstrates the presence of specific receptors for VIP in VSMCs functionally coupled to adenylate cyclase system. It is suggested that VIP exerts its vasodilatory effect through its specific receptors distinct from β-adrenergic receptors.     

Agonist binding promotes a guanine nucleotide reversible increase in the apparent size of the bovine anterior pituitary dopamine receptors

Dopamine receptors, solubilized from bovine anterior pituitary membranes with the detergent digitonin, retained a typical dopaminergic specificity for the binding of both agonists and antagonists. The affinities of antagonists for binding to the soluble receptors are virtually identical with those observed with the membrane-bound receptors. The affinities of agonists however, correspond to those for the form of the receptors in the membranes having low affinity for those agonists (De Lean, A., Kilpatrick, B. F., and Caron, M. G. (1982) Mol. Pharmacol. 22, 290-297). Thus, after solubilization, agonist high affinity interactions with the receptor and their sensitivity to modulation by guanine nucleotides are lost. However, high affinity agonist binding and its sensitivity to guanine nucleotides can be preserved if the membrane-bound receptors are prelabeled with the agonist [3H]n-propylapomorphine prior to solubilization. In order to investigate the molecular basis for these changes in the properties of agonist binding, the solubilized receptors were characterized by chromatographic procedures. Using molecular exclusion high pressure liquid chromatography, [3H]n-propylapomorphine-prelabeled receptors elute as an apparent larger molecular species than either unlabeled or antagonist [( 3H]spiroperidol)-pre-labeled receptors. Moreover, incubation of the pooled agonist-prelabeled receptor peak with guanine nucleotides effects a decrease in the apparent size of the receptors such that upon rechromatography they elute in a position coincidental with the 3H-antagonist-pre-labeled receptor peak. Thus, occupancy of the receptors by agonists promotes the formation of a guanine nucleotide-sensitive agonist high affinity form of the receptor which is of larger apparent size presumably due to the association of the receptor with a guanine nucleotide regulatory protein.     

High-affinity receptors for vasoactive intestinal peptide on human myeloma cells

Cultured human myeloma cells of the U266 line and leukemic T cells of the Jurkat line bound synthetic [125I]Tyr10-vasoactive intestinal peptide1-28 ([125I]VIP1-28) specifically and with an affinity similar to that of neuroendocrine cells. Specific binding reached equilibrium after 2 h at 22 degrees C for both myeloma cells and T cells, attained a maximum of 57 to 71% of total binding, and was reversed in 1.5 to 3 h by an excess of non-radioactive VIP1-28. Analyses of the ligand concentration-dependence of binding of the ligand concentration-dependence of binding of [125I]VIP1-28 revealed a mean Kd of 7.6 nM for a mean of 41,207 receptors per myeloma cell and 5.2 nM for 12,266 receptors per T cell. The relative affinity of binding of mast cell-derived VIP10-28 free acid and synthetic analogues suggested differences in specificity between lymphocyte and neuroendocrine receptors. Distinct sets of receptors thus appear to mediate the effects of VIP on functions of both antibody-producing cells and T cells.     

Characterization of functional receptors for vasoactive intestinal peptide (VIP) in rat peritoneal macrophages.

Functional vasoactive intestinal peptide (VIP) receptors have been characterized in rat peritoneal macrophages. The binding depended on time, temperature and pH, and was reversible, saturable and specific. Scatchard analysis of binding data suggested the presence of two classes of binding sites: a class with high affinity (kd = 1.1 +/- 0.1 nM) and low capacity (11.1 +/- 1.5 fmol/10(6) cells), and a class with low affinity (kd = 71.6 +/- 10.2 nM) and high capacity (419.0 +/- 80.0 fmol/10(6) cells). Structural requirements of these receptors were studied with peptides structurally or not structurally related to VIP. Several peptides inhibited 125I-VIP binding to rat peritoneal macrophages with the following order of potency: VIP greater than rGRF greater than hGRF greater than PHI greater than secretin. Glucagon, insulin, somatostatin, pancreastatin and octapeptide of cholecystokinin (CCK 26-33) were ineffective. VIP induced an increase of cyclic AMP production. Half-maximal stimulation (ED50) was observed at 1.2 +/- 0.5 nM VIP, and maximal stimulation (3-fold above basal levels) was obtained between 0.1-1 microM. Properties of these binding sites strongly support the concept that VIP could behave as regulatory peptide on the macrophage function.     

Presence of vasoactive intestinal peptide receptors coupled to adenylate cyclase in rat lung membranes

(1) The binding of 125I-labelled vasoactive intestinal peptide (VIP) to a particulate fraction from rat lung was rapid, temperature dependent, saturable and specific. This process was also reversible and 125I-labelled VIP dissociation was accelerated by guanine triphosphate nucleotides. The curves describing the inhibition of tracer binding by peptides of the VIP-secretin family suggested the presence of at least two classes of VIP receptor: a "high-affinity' type with decreasing affinity for VIP in the order: VIP = [Val5]secretin greater than [Ala4, Val5]secretin; and a "low-affinity type' with decreasing affinity for VIP in the order: VIP greater than [Val5]secretin greater than [Ala4, Val5]secretin = secretin greater than [Ala4]secretin. (2) VIP and related peptides stimulated the adenylate cyclase activity of the same lung membrane preparation more efficiently than beta-adrenergic agonists and prostaglandins E1 and E2. The dose-effect curves of stimulation of adenylate cyclase by VIP and parent peptides were also compatible with the existence of two classes of VIP receptor, the relative peptide potencies being identical with their ability to compete with 125I-labelled VIP for binding.     

Characterization of vasoactive intestinal peptide receptors by a photoaffinity label. Site-specific modification of vasoactive intestinal peptide by derivatization of the receptor-bound peptide

The biological effects of vasoactive intestinal peptide (VIP) are mediated by binding to a membrane-bound receptor. Probes designed to trap this receptor by binding to it in a covalent way may suffer from a greatly reduced affinity. We report here, for the VIP receptor, the use of a photoaffinity probe obtained by derivatization of receptor-bound VIP with para-azidophenylglyoxal. This method protected the parts of the molecule essential for receptor binding. The VIP derivative thus obtained became covalently linked when irradiated. In the dark, however, it exhibited normal VIP-like behavior and retained its biological activity. This derivatization method might be generally applicable when hormone analogues have to be prepared without loss of receptor affinity. Receptor characterization studies on liver plasma membranes showed the presence of high- and low-affinity binding sites with KD = 0.1 and 5 nM, respectively. Treatment of membranes with dithiothreitol causes loss of high-affinity binding. The high-affinity site, trapped by the photoaffinity probe, resolved into two molecular mass forms, 50 and 200-250 kDa. Reduction of the receptor-probe complex left the 50-kDa form intact, whereas the amount of the 200-250-kDa form greatly diminished. We demonstrate the importance of the presence of disulfide bonds in one of the molecular forms involved in high-affinity binding.     

Importance of disulfide bonds in receptors for vasoactive intestinal peptide and secretin in rat pancreatic plasma membranes

(1) Vasoactive intestinal peptide (VIP), secretin, and C-terminal octapeptide of cholecystokinin (CCK-8) receptors were identified in rat pancreatic plasma membranes by the ability of these peptides to stimulate adenylate cyclase activity. The membrane preparation procedure was conducted through a series of steps including discontinuous sucrose density gradient fractionation. 5 mM β-mercaptoethanol was added stepwise. Membrane preparations obtained stepwise were preincubated for 10 min at 25°C in the presence of various concentrations of β-mercaptoethanol or dithiothreitol before assaying adenylate cyclase. The use of the reducing agents exerted no effect on p[NH]ppG-, NaF-, and CCK-8- stimulated activities. By contrast, stimulation of adenylate cyclase by low VIP concentrations was specifically altered when β-mercaptoethanol was used during tissue homogeneization at 5°C. (2) In addition, both VIP and secretin responses were highly sensitive towards a preincubation of 10 min at 25°C in the presence of dithiothreitol. (3) These results were likely to reflect alterations at the receptor level. ¹²⁵I-VIP binding was, indeed, reduced after dithiothreitol preincubation, low concentrations of the thiol reagent decreasing the apparent number of high-affinity VIP receptors and higher dithiothreitol concentrations reducing the affinity of VIP receptors.     

Isolation and characterization of rat vasoactive intestinal peptide

We have used gel filtration, ion exchange chromatography, affinity chromatography and reversed-phase HPLC to isolate vasoactive intestinal peptide from rat intestine. Microsequence analysis of 1 nmole peptide indicated that the sequence was identical to the porcine octacosapeptide VIP. In radioimmunoassay with four antisera and in the turkey pancreas bioassay, rat VIP was equipotent with highly purified preparations of porcine, human and canine VIP. A less basic rat VIP-variant was also isolated and the N-terminal decapeptide region that was sequenced was identical with that of porcine VIP.     

Evidence for vasoactive intestinal peptide receptors in apical membranes from tracheal epithelium

[125I]VIP (vasoactive intestinal peptide) bound to apical membranes isolated from the bovine tracheal epithelium with a half maximal inhibition by unlabeled VIP (IC50) of 0.6 x 10(-9)M and binding was reversible. Glucagon did not affect [125I]VIP binding to the membranes. [125I]VIP was covalently cross-linked to tracheal membrane proteins using disuccinimidyl suberate. SDS-polyacrylamide gel electrophoresis of labeled tracheal membranes revealed one major [125I]-receptor complex of Mr = 71,000 to which binding of [125I]VIP was inhibited by 10 microM unlabeled VIP. These results are consistent with the presence of a specific, high-affinity receptor for VIP, with a Mr = 71,000, in apical membrane vesicles isolated from the bovine tracheal epithelium.     

Early changes in pituitary adenylate cyclase-activating peptide, vasoactive intestinal peptide and related receptors expression in retina of streptozotocin-induced diabetic rats

The retinal expression and distribution of pituitary adenylate cyclase-activating peptide (PACAP) and vasoactive intestinal peptide (VIP) and their receptors was investigated in early streptozotocin (STZ)-induced diabetic rats. Diabetes was induced in rats by STZ injection (60mg/kg i.p.). PACAP, VIP and their receptors in nondiabetic control and diabetic retinas were assayed by quantitative real-time PCR and Western blot 1 and 3 weeks after STZ injection. Effects of intravitreal treatment with PACAP38 on the expression of the two apoptotic-related genes Bcl-2 and p53 were also evaluated. PACAP and VIP, as well as VPAC1 and VPAC2 receptors, but not PAC1 mRNA levels, were transiently induced in retinas 1week following STZ. These findings were confirmed by immunoblot analyses. Three weeks after the induction of diabetes, significant decreases in the expression of peptides and their receptors were observed, Bcl-2 expression decreased and p53 expression increased. Intravitreal injection of PACAP38 restored STZ-induced changes in retinal Bcl-2 and p53 expression to nondiabetic levels. The initial upregulation of PACAP, VIP and related receptors and the subsequent downregulation in retina of diabetic rats along with the protective effects of PACAP38 treatment, suggest a role for both peptides in the pathogenesis of diabetic retinopathy.     

Identification and characterization of five-transmembrane isoforms of human vasoactive intestinal peptide and pituitary adenylate cyclase-activating polypeptide receptors

The seven-transmembrane (7TM) G-protein-coupled neuroendocrine receptors VPAC1 (HGNC approved gene symbol VIPR1) and VPAC2 (HGNC approved gene symbol VIPR2) are expressed in different tissues and involved in the regulation of important biological functions. We now report the identification and characterization of novel five-transmembrane(5TM) forms of both human VPAC1 and human VPAC2. These alternatively spliced variant mRNAs result from the skipping of exons 10/11, spanning the third intracellular loop, the fourth extracellular loop, and the transmembrane regions 6 and 7, producing in-frame 5TM receptors predicted to lack a G-protein-binding motif. RT-PCR showed that these 5TM receptors are differentially expressed in transformed and normal cells. Translation of the 5TM protein was demonstrated by transfection and expression in CHO cells. Following agonist stimulation, differential signaling of the 7TM versus 5TM forms was shown both for the activation of adenylate cyclase and for tyrosine phosphorylation. The identification of these splice variants in various cells and their expression and differential signal transduction compared to the 7TM form suggest that these novel receptors have biological relevance.     

Alpha-melanocyte-stimulating hormone secretion from permeabilized intermediate lobe cells of rat pituitary gland. The role of guanine nucleotides

The non-hydrolyzable GTP analogue, guanosine 5'-O-(3-thiotriphosphate) (GTP gamma S) and cyclic AMP potentiated the Ca2+-evoked secretion of alpha-melanocyte-stimulating hormone (alpha-MSH) from permeabilized neurointermediate lobe (IL) cells of rat pituitary gland. The enhancement by Mg-GTP gamma S (100 microM) and cyclic AMP (1 microM) depended on the intracellular Ca2+ concentration (EC50 = 4.8 +/- 1.8 and 4.6 +/- 1.7 microM; mean +/- SE, with and without Mg-GTP gamma S and cyclic AMP, respectively). A similar effect was observed with guanine nucleotide triphosphate (GTP and GppNHp). Mg was absolutely required for this event. Neither Mg-GTP gamma S nor cyclic AMP alone was effective in potentiating alpha-MSH secretion. GDP beta S blocked the Mg-GTP gamma S (100 microM) and cyclic AMP augmented secretion of alpha-MSH. Neither neomycin (which affects the process of inositol 1,4,5-triphosphate-mediated Ca2+ mobilization) or colchicine (which influences microtubule assembly) had an effect on the cyclic AMP and Mg-GTP gamma S potentiation of alpha-MSH secretion. These data suggest that the GTP-binding protein may be involved in the regulation of alpha-MSH secretion after Ca2+ entry into the cells, since the intracellular environment is controlled in the permeabilized cells.     

Identification of functional receptors for vasoactive intestinal peptide and neurotensin in the human submandibular gland duct cell line, HSG-PA.

The HSG-PA human submandibular gland adenocarcinoma cell line has attracted attention recently as a potentially useful cell culture model for studies of salivary duct cell function and regulation. These cells possess a variety of morphological and biochemical markers found in salivary duct cells. Recently, muscarinic cholinergic receptors coupled to inositol intracellular Ca2+ mobilization (He et al., Eur. J. Physiol., 413 (1989) 505-510) and K+ fluxes (Ship et al., Am. J. Physiol., 259 (1990) C340-C348) have been identified in HSG-PA cells. In this study, we report the presence in these cells of functional receptors for two neuropeptides, vasoactive intestinal peptide (VIP) and neurotensin. Receptors for both peptides were labeled in intact cell radioligand binding studies and exhibited pharmacological profiles similar to receptors found in other tissues. There was close agreement between binding Ki values and the ED50 values for stimulation of second messenger production and modulation of K+ efflux, with all values between 1 and 5 nM. Whereas neurotensin stimulated K+ efflux dramatically, VIP alone had no effect but enhanced the response to neurotensin. These studies thus represent the initial documentation of functional receptors for VIP and neurotensin in a cell line of salivary duct cell origin.