Multiple signaling pathways for EDG (endothelial differentiation gene) receptors in gastric smooth muscle

604

MARCKS is regulated by PKC. These findings provide novel information regarding signaling mechanisms contributing to stimulated gut peptide secretion.

A Comparative Study of Patient Perceptions and Screening Preferences for StoolBased DNA Testing (SBDNA), Fecal Occult Blood Testing (FOBT) or Colonoseopy (CS) Paul C. Schroy III, Timothy C. Heeren

607 Signal Transduction Pathways Regulating Parietal Cell Dedifferentiation Vinzenz Stepan, Saravanan Ramamoorthy, Hildegard Nitscbe, Kilsti Brown, Andrea Todisco

Background: Poor patient acceptance is a major barrier to effective colorectal cancer (CRC) greening. SBDNA is a novel CRC screening strategy that offers a convenient, non-invasive and potentially more acceptable alternative to existing screening tests. The objective of this studywas to indirectly assess patient acceptance by comparing patient perceptions of SBDNA, FOBT and CS and eliciting preferences. Methods: A prospective survey was conducted amongpreviously unscreened, asymptomatic average-risk subjects 50 years of age and older panicipating in a multi-center comparison of SBDNA (PreGen-PhisTM)and FOBT (Hemoccult If) for detecting CRC neoplasia. Evahiable subjects completed a 25-item questionnaire within 48hrs after undergoing a screening colonoscopy, which served as the reference standard for SBDNAand FOBT test performance. Respondents were asked to rate each of the 3 screening tests on a variety of features using a 5-point ordinal scale and select a preferred strategy. Results:A total of 2,388 subjects completed the survey. SBDNA received higher mean ratings (p<0.001) than FOBT for simplicity of the stool sample collection process, comfort, perceived accuracyand likelihood of repeating the test if recommended by their doctor. SBDNA and FOBT received similar ratings for simplicity of prep instructions, lost time from work or otheractivities, perceived invasiveness, perceived embarrassment and perceived anxiety over the prep and test itself. FOBT was never rated higher than SBDNA with respect to any feature. When compared with CS, SBDNA received higher ratings (p<~0.001) for all of the alorementioned test features except accuracy, where CS was rated higher (p<~O.001). Overall, a higher percentage (p<0.001) of patients preferred SBDNA (50.1%) to both FOBT (34.6%) and CS (15.3%) for their routine screening program. Conclusions: Average risk patients perceive SBDNA to have a number of advantages over both FOBT and CS. Moreover, our data suggest that SBDNA may be a preferred strategy for routine CRC screening.

The Sonic Hedgehog (Shh) signal transduction pathway regulates cellular growth and differentiation. Shh binds to the receptor protein patched (Ptc), relieving the tonic inhibitory effect of Ptr on the transmembrane protein Smoothened (Smo). This allows the activation of downstream targets through the Gli family of transcription factors. We previously reported that prolonged incubation (> 72 h) of purified (>95%) canine gastric parietal cells in primary culture with EGF induces parietal cell dedifferentiation, a process characterized by morphological transformation, inhibition of H+/K+-ATP-ase gene expression and induction of cell proliferation. In this study we explored the role of Shh/Ptc/Smo in the dedifferentiating actions of EGF in the parietal ceils. EGF (10 nM) suppressed the expression of Shh and Ptc but~hot of Smo in the parietal cells after 72~h0 f i ~ b a t i o n . Shh, Ptc and Smo expression were measured by western blots and by immuno~bchemieal staining of the parietal ceils with anti-Shh, -Ptc and -Smo antibodies. We examined the consequences of loss of expression of the inhibitory protein Ptc. Cyclopamine (1jzM) an inhibitor of Smo signaling, reversed EGF-induced parietal cell morphological transformation and it inhibited EGF induction of both cyclin DI and TGF-c~ expression in the parietal cells, measured by westem blots with anti-cyclin D 1 and -TGFa antibodies. Cyclopamine also inhibited MAPK activation, measured by western blots with anti-phospho-MAPK antibodies and it blocked EGF-stimulated parietal cell proliferation, assessed by incorporation of tritiated thymidine into the DNA. To further explore if deregulated Smo signaling affects TGFa expression, we co-transfected the parietal cells with a hiciferase reporter plasmid containing 2813 bases of the human TGFct gene promoter together with a dominant negative Gh2 gene. EGF induced TGFa luciferase activity two-fold, and dominant negative Gli2 inhibited this effect. Since Akt promotes parietal cell differentiation, we examined the effect of Akt on Shh and Ptc expression in EGF-treated, dedifferentiated parietal cells. Adenovirus mediated gene transfer of constitutively active Akt into the dedifferentiated parietal cells, restored the expression of both Shh and Ptc. Accordingly, prolonged exposure of the parietal cells to EGF leads to parietal cell dedifferentiation through loss of Ptc and continuous, deregulated signaling from Smo. Akt appears to be a crucial switch for the activation of programs of parietal cell differentiation.

605

The Extracellular Domain of Receptor Activity-Modifying Protein-1 (RAMP-l) is Sufficient for Calcitonin Receptor-Like Receptor (CRLR) Function Timothy]. Fitzsimmons, Xilin Zhao, Stephen A. Wank Background:Calcitonin Gene-Related Peptide (CGRP) is a neuropeptide found in the central andperipheral nervous systems. CGRE mediates sensory neurotransmission, decreases vascular tone, gastrointestinal motility and secretion, and inhibits the action of insulin on carbohydrate metabolism. A functional CGRP receptor requires dimerization of Calcitonin Receptorhke Receptor (CRLR) with Receptor Activity-Modifying Protein-1 (RAMP-I). Purpose: To determine the function of the three domains (extracelhilar (ECD), transmembrane (TM), and tail domains) of human RAMP-1. Methods: Three mutants were constructed: 1) RAMP1 without the cytoplasmic tail, 2) a chimera consisting of the ECD of RAMP-1 and the TM and tail of the Pfatelet Derived Growth Factor Receptor (PDGFR) and 3) the ECD of RAMP1 alone. These RAMP-1 mutants were examined for their ability to associate with CRLR to effectCGRP stimulated cAMP accumulation and esI-CGRP binding, as well as CRLR trafficking and cell surface expression as determined by flow cytometry and immunoblotting. Results: All RAMP-1 mutants were able to associate with CRLR with full efficacy for CGRP stimulated cAMP accumulation. However, the RAMP-1/PDGFR chimera demonstrated a 10f01ddecrease in potency for CGRP signaling and a similar decrease for CGRP binding. The RAMP-1 ECD mutant had a 4000-fold decrease in potency and binding was too weak to measure. CRLR cell surface expression was retained with all RAMP-1 mutants, although the RAMP-1 mutant consisting of the ECD alone functioned less efficiently and was secreted into the medium. Conclusions: The ECD of RAMP- 1 is sufficient for normal CRLR association and efficacy. The presence of a TM domain and the specific sequence of the RAMP-1 TM domaincontribute to CGRP affinity and potency. The C-terminal tail of RAMP- 1 is unnecessary for CRLR function.

6O8 Cooperation of GQ, GI and G12/13 in Proteig/rl~tnase D Activation and Phosphorylation Induced by Lysophosphattdie Acid Jingahen Ynan, Lee W. Slice, Jennifer Gu, Enrique l~zengurt STUDY PURPOSE: Lysophosphatidic acid (LPA) promotes a broad range of biological responses and multiple molecular events in target cells. Consistent with the stimulation of multiple signaling pathways, LPA has been shown to activate severalheterotrimeric G proteins including Gq, Gi and G13 in Swiss 3T3 cells. Protein kinase D (PKD/PKCIz) is a seilne/ threonine protein kinase with distinct structural, enzymological and regulatory properties. Recently, activation of a number of receptors that couple to heterotilmeric G proteins including those for LPA has been shown to stimulate PKD activation in a variety of cell types. PKC-dependent PKD activation has been identified as an early event in the action of LPA in intact Swiss 3T3, Rat-1 and IEC-6 cells. In order to examine the contribution of different G-protein pathways to LPA-induced PKD activation, we tested the effect of LPA on PKD activity in mutine embryonic cell lines deficient in Gaq/11 (Guq/11 KO cells) or G~ 12/13 (Get12/13 KO cells), and used cells lacking rhodopsin kinase (RIg cells) as a control. RESULTS: In RK and Gut12/13 KO cells, LPA induced PKD activation through a PLC/PKC pathway in a concentration-dependent fashion with mammal stimulation (6-fold for RK cells and 4-fold for Ga12/13 KO cells in autophosphorylation activity) achieved at 3 ~M. In contrast, LPA did not induce any significant increase in PKD activity in Gaq/11 KO cells. However, LPA induced a significantly increased PKD activity when Gaq/11 KO cells were transfected with Gctq. LPA-induced PKD activation was modestly attenuated by prior exposure of RK ceils to pertussis toxin (Fix) but almost abolished by the combination treatments of Fix and Clostridium difficile toxin B, a potent inhibitor of Rho. Surprisingly, Fix alone strikingly inhibited LPA-induced PKD activation in a concentration-dependent fashion in Ga12/13 KO cells. Similar results were obtained when activation loop phosphoryfation at Ser-744 was determined using a specific antibody that detects the phosphorylated state of this residue. CONCLUSION: Our results indicate that Gq is necessary but not sufficient to mediate LPA-indnced PKD activation. In addition to Gq, LPA requires additional G-protein pathways to elicit a maximal response with Gi playing a critical role in Ga12/13 KO cells. We conclude that LPA induces PKD activation through Gq, Gi and G12, and propose that PKD activation is a point of convergence in the action of multiple G-protein pathways.

606 Neurotensin Release In Response To Protein Kinase C Activation By Phorbol Esters Is Mediated Through Myristoylated Alanine-Rich C Kinase Substrate Phosphorylation Jiag Li, Kathleen L. O'Connor, Leoncio Vergara, Mark R. Hellmich, George H. Greeley Jr, Courtney M Townsend Jr., B. Mark Evers Myristoylated alanine-rich C-hnase substrate (MARCKS) is a cellular substrate for protein kinase C (PKC). Recently, we have shown that PKCa and B play a role in pborbol 12myristate 13-acetate (PMA)-mediated secretion of neurotensin (NT), a gut peptide which affects GI secretion, motility and growth. In our present study, we determined whether PKC-mediated NT secretion is due to the activation of MARCKS using the novel endocrine cell line BON, which produces and secretes NT peptide. METHODS. To determine the role 0f MARCKSin PKC-mediated NT secretion, stable transfection of bovine MARCKS, (wildtype, myristoylation and phosphorylation site domain (PSD) mutants), was performed in BON cells; NT secretion was measured by RIA. Transloeation of GFP-tagged MARCKS or endogenous MARCKS was observed by time series confocal analysis or subcellular fractionation, respectively. Phosphorylation was assessed by Westem blots. RESULTS. Overexpression of wild-type MARCKS resulted in a significant increase in PMA-mediated NT secretion compared with the empty (ie, control) vector. GFP-tagged wild-type and PSD mutant MARCKSproteins were expressed in the membrane but the myristoylation site mutant was detected only in the cytosol. In addition, translocation was detected from membrane to cytosolby time series confocal analysis lmin after PMA addition only in wild-type MARCKS in living cells, demonstratmg the function of MARCKS is myristoylation and PSD phosphorylation dependent. Pretreatment with PKC inhibitors (Go6983, GFX and Ro31-8220) effectivelyinhibited the phosphorylation of MARCKS, suggesting the regulation of PKC on the activity of MARCKS. CONCLUSIONS. We demonstrate, for the first time, that MARCKS protein is involved in PMA-mediated NT secretion in BON ceils. Furthermore, activity of

609 Multiple Signaling Pathways for EDG (endothelial differentiation gene) Receptors in Gastric Smooth Muscle Huipmg Zhou, Karnam S. Murthy, Sankar Das, John R. Gilder, Gabriel M. Makhlouf Eight EDG receptors have been cloned, five of which (EDG 1, -3, -5, -6, and 8) are selectively activated by sphingosine-l-phosphate (SIP). We have recently shown that EDG5 receptors are expressed in gastric smooth muscle cells. In other cells types, EDG receptors couple variously to Fix-sensitive and -insensitive G proteins. Aim: To characterize the signahng pathways initiated by S1P in rabbit gastric muscle cells. Methods: G protein activation was determined in freshly dispersed muscle cells. PI hydrolysis and cAMP formation were measured in freshly dispersed and cultured muscle cells and in cultured muscles cells transfected with various Ga minigenes. Muscle contraction was measured by scanning micrometry. Results: SIP stimulated the activities of Gq, G~3, and all three isoforms of G,,

A-77

AGA

Abstracts

but not G,. and induced PI hydrolysis. The latter was abolished by U73122, and was partly inhibited by pertussis toxin (PTx), GI3 antibody, and Gaq antibody. A combination of Gaq and G[3 antibodies abolished SIP-stimulated PI hydrolysis, implying that PI hydrolysis was mediated by Gaq and G~3%. In support of this notion, PI hydrolysis was partly inhibited by expression of Gq or G, minigenes. S1P inhibited forskolin-stimulated cAMP formation, which was abolished by pretreatment of the cells with PTx or by expression of G~ minigene. S1P induced a concentration-dependent contraction (ECs0:5 nM). Contraction consisted of an initial transient phase followed by a sustained phase. The initial phase was abolished by U73122 and the MLCK inhibitor, ML-9, and was partly inhibited by PTx and by antibodies to Gc% and GI3. A combination of the two antibodies abolished initial contraction. In contrast, the sustained phase of contraction was partly inhibited by the PKC inhibitor, bislndolylmaleimide, and the Rho kinase inhibitor, Y27632, and was abolished by a combination of the both inhibitors. Sustained contraction was partly inhibited by Gcr 3 and G% antibodies and was abolished by a combination of both antibodies, implying that the PKC/ Rho kinase-dependent pathways were downstream of both Gq and GI3. Consistent with this notion, Rho kinase activity was inhibited by expression of Gq minigene. PTx had no effect on sustained contraction implying that this phase was not dependent on G~activity. Conclusion: EDG receptors activate Gq, G,, and G~3 in gastric muscle. Gq and G, mediate initial contraction by stimulating PI hydrolysis, whereas Gq and G~3 mediate sustained contraction by activating RhoA.

612

Eating Frequency and Colon Cancer Risk Jeffrey T. Wei, Alexandra Connelly, Jessie S. Abouta, Joseph A. Galanko, Christopher Martin, Robert S. Sandier PURPOSE: Previous studies have suggested an association between increased eating frequency and colon cancer. Two potential mechanisms may underlie this association: 1) increased exposure of the colon to oncogenic, secondary bile acids, or 2) elevation in insulin or insulinlike growth factor, which promote in vitro proliferation of colon cancer cells. The objective of this study was to further assess this relationship in a population-based study. METHODS: Data were obtained from a population-based case-control study of incident colon cancer in 33 counties of North Carolina from 1997-2000. Population-based controls were derived from the Health Care Financing Administration for subjects 65 years and older, and from the Department of Motor Vehicle tapes for subjects less than 65 years of age. The present analysis focused on questionnaire items related to frequency of eating. The number of meals, and number of snacks per day were reported by both case and control subjects. Eating frequency (a combination of meals and snacks) was categorized as <2, 3-4, or >5 eating episodes/day. Multivariate logistic regression analysis was performed to calculate odds ratios for the association between meal frequency and colon cancer, adjusting for possible confounders. Coffee intake, sex, BMI, carbohydrate intake, and physical activity were also evaluated as effect modifiers. RESULTS: 313 subjects with colon cancer and 590 control subjects were eligible for analysis. Sex interacted with the relationship between eating frequency and colon cancer. Women in the highest category of eating frequency had a similar risk of cancer as women in the lowest category (adjusted OR 0.64, 95%CI 0.34-1.21). In contrast, men in the highest category were at more than twice the risk compared to men in the lowest category (adjusted OR 2.30, 95% CI 1.04-5.06). Coffee intake, BMI, carbohydrate intake, and physical activity did not interact with this relationship. CONCLUSIONS: Increased eating frequency was associated with colon cancer among men, but not women. Sex differences in bile acid metabolism may be responsible for this difference.

610 Mechanism of JAK2 Activation by the G Protein Coupled Receptor CCK2-R Audrey Ferrand, Aline Chauve[, Chantal Escrieut, Jean-Pierre Esteve, Lucien Pradayrol, Daniel Fourmy, Marlene Dufresne, Catherine Seva Numerous studies show that gastrin is a growth factor for the GI tract: Ins-GAS mice, overexpressing gastrin, have a gastric hypertrophy and develop gastric cancer. In addition, gasmn proliferative effects have been shown on pancreatic, gastric and colon tumoral cells. This hormone acts via a G-protein coupled receptor (GPCR), the CCK2-receptor (CCK2R). Our aim is to identify molecular mechanisms linking the CCK2-R to pathways involved in proliferation. Janus kinases induce cell growth in response to certain cytokines or growth factors, but the role of these tyrosine-kinases in GPCR signaling remains largely unknown. Among JAK members, JAK2 is known to be involved in cell proliferation and cancer. In AR4-2J ceils expressing endogenous CCK2-R, we detected a rapid and transient activation of JAK2 in response to gastrin. To study the mechanism of JAK2 activation by CCK2-R we used COS-7 cells co-transfected with JAK2 and CCK2-R. By immunoprecipitations and western-blot analysis, we found a constitutive association between JAK2 and the CCK2-R. We tested different hypothesis for the direct or indirect interaction of JAK2 with the receptor: 1) Tyrosine of the E/DRY motif in the chemokine receptor, CCR21% is known to interact with JAK2. However, point mutation of this motif in the CCK2-R (Y153F) did not affect the association with JAK2 or its activation. 2) Similar results were obtained with mutation of proline rich sequence in the C-terminal tail of the CCK2-R in contrast to what was observed for cytokines receptors. 3) Previous studies have shown that SHP2 may play an adaptator role between JAK2 and the ATl-receptor. Using Surface Plasmon Resonance technique, we found that the phosphorylated tyrosine 438 of CCK2-R permits the direct association of SHP2 to the receptor. SHP2 can also associate with JAK2. However, the receptor mutant Y438F associates and activates JAK2 in COS-7 cells. 4) We previously shown that Ash-391 in the NPXXY motif of the CCK2-R is involved in Gq proteins activation. Indeed, the N391A receptor mutant is unable to mediate pathways downstream of Gq such as PLC. We found that this mutation inhibits JAK2 activation in response to gastrin. In conclusion, gastrin induces JAK2 activation via CCK2-R. The receptor and the tyrosinekinase are constitutively associated. Moreover, we show that the NPXXY motif in the receptor sequence and probably G alpha q are involved in JAK2 activation. To our knowledge, it is the first time that the NPXXY domain is shown involved in JAK activation.

613 Insulin, Insulin-like Growth Factor Proteins, and Colorectal Cancer Risk in the ATBC Study Paul Limburg, Robert Vierkant, Rachael Stolzenberg-Solomon, Thomas Sellers, Michael Pollak, Jarmo Virtamo, Demetrius Albanes Background: Insulin and insulin-like growth factor (1GF) proteins may be associated with colorectal cancer (CRC) risk. However, there are few prospective data and the role of insulin resistance, as estimated by homeostasis model assessment (HOMA-IR), is unknown. Aim: To evaluate associations between insulin, IGF-1, IGF binding protein-3 (IGFBP3), HOMAIR and incident CRC in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention (ATBC) Study. Methods: CRC cases (n = 142) and cancer-free controls (n = 400) were selected from among 29,133 male Finnish smokers enrolled in the ATBC Study using a nested case-control design. Questionnaire data and serum samples were obtained prior to randomization (from 1985-88) Case status was assessed for up to 12 years. Odds ratios and 95% confidence intervals (OR; 95% CI) were estimated using multivariate logistic regression models. Results: Cases were slightly older than controls (mean age at baseline 59.1 and 56.4 years, respectively; p<0.001). Other demographic, dietary, and lifestyle factors were not significantly associated with case status, except physical activity at work (inverse association; p=0.001). Risk estimates for insulin, IGF-1, IGFBP3, IGF-1/IGFBP3 ratio, and HOMA-IR are shown in the table. Conclusions: Prediagnostic measures of insulin, IGFBP3, and HOMA-IR were not significantly associated with colon cancer, rectal cancer, or CRC risk. Men in the highest vs. lowest quartiles of IGF-1 and IGF-1/IGFBP3 bad significantly reduced risks for rectal cancer and CRC, respectively, but the linear trend tests were only nominally statistically significant. Our findings differ from some pnor reports, which suggests that cigarette smoking may influence 1GF protein-related CRC risks. This hypothesis requires further investigation.

611 NSAID use, gender, age and other predictors of proximal neoplasia in a screening population Joseph C. Anderson, Zvi Alpern, Bernard Lane, Patricia Ells, Peter F. Ells, Douglas L. Brand, Roger Grimson

AGA Abstracts

p 0.0001 NS NS

~ neoplalil 0R=2.9(05%CI;2.0-4.0) OR=,64(95%CI;.46...80) OR=1.6(95%CI;1.1-2.4)

p 0.0001 0.009 0,001

OR=.53(95%CI;.28-.98}

0.04

NS

NS

OR (95% Cl)=

OR (95% Cly

p trend

p trend

Colorectal Cancer {n=142)1 OR (95% Ci)=

p trend

0,99 (0.460.69 1.10 (0.400,83 1.05 (0.56. 0.76 2.11) 3.04) 1.98) IGF.t 1,23 (0.470.70 0,25 (0.050.10 0.83 (0.380.69 3.26) 0.99) 1.80) IGFBI=3 1.57(0.610.24 0,61 (0.160.57 1.10 (0.510.66 4.05) 2.24) 2.38) IGF-I/IGFBP3 0.55 (0.260.20 0.47 (0.170.13 0.52 (0.290.07 1.15) 2.16) 0,95) HOMA.IR 1,05 (0,550.87 1,00 (0.42. 0.69 1.20 (0200.91 2.04) 2.38) 2,05) 1Includes 12 cases with no specified subsite. 2Quartile 4 vs. quartJle1, adjusted for age at baseline. intervention group, physical activity at work, and other serum analytes (glucose, insulin, IGFt, IGFBP3) as appropriate.

614 The Effect Of Histamine Receptor Antagonists on Incident Colorectal Adenomas Brian J. Schwender, Douglas J. Robertson, Carol A. Burke, Leila Mort, John A. Baron, for the Polyp Prevention Study Group

Risk Factom of Significant or Any Proximal Colorectal Neoplella (Odds Ratio) Signiflcaat neOl)lalda OR--4.0(95%01;2.2.7.2) NS NS

Rectal Cancer (nm46)

Insulin

Background: Knowledge regarding risk factors for proximal colorectal neoplasia is important especially in the debate of flexible sigmoidoscopy versus colonoscopy. Methods: 1864 patients who bad a screening colonoscopy were assessed for age, gender, family history, BMI, education, ethnicity, alcohol use, smoking, physical activity, fruit/vegetable intake. NSAID users were defined as those who took at least an aspirin (325 rag) or an NSAID daily for ten years(90% took ASA). Splenic flexure divided lesions into distal or proximal. Significant neoplasia was defined as villous, greater than one cm, high grade dysplasia, multiple (> 2) or adenocarcinoma. Any adenomatous neoplasia was defined as significant (see above) or single tubular adenomas. Multivariate analysis was done examining the impact of all risk factors on two endpoints: any adenomatous and significant neoplasia. Results/Conclusions: Overall 3.4 % of the patients had significant and 10.0% had any adenomatous pathology. Of note, NSAIDS did not reduce left sided neoplasia. Age and NSA1D use influenced proximal significant neoplasia, while age, gender and family history influenced any adenomatous neoplasia. NSAIDS were protective of right but not left sided significant neoplasia.

Risk Factor Age>57 Fentale Getcinr Family hlstow of r NSND use

Colon Cancer (r1,=84)

Background: We previously examined the effect of histamine receptor antagonists on adenoma risk in a randomized chemoprevention trial and found suggestions that H2 receptor antagonists (H2RA's) may reduce risk of incident adenomas. We now report the role of H1 receptor antagonists (H1RA's) and H2RA's in two subsequent chemoprevention trials. Methods: We determined the risk of adenoma recurrence in relation to H1RA and H2RA exposure among subjects in the calcium (CPPS) and the aspirin/folate polyp prevention studies (APPS). Subjects with adenomas were identified at baseline and randomized to study drug (i.e

A-78