AG

AA

bst

ract

sthrough Matrigel compared with SW480 cells transfected with non-targeting control shRNA.CONCLUSIONS. Knockdown of AKT1 in SW480 cells induces EMT formation and pro-motes cell invasion by downregulating and disrupting E-cadherin expression. Since AKT2has been implicated in the promotion of CRC cell metastasis, our current study suggeststhat AKT1 and AKT2 play distinct roles in regulating CRC cell metastasis, thus providinga better understanding of the important and distinct functions of AKT isoforms duringCRC progression.

W1889

CD133 and Side Population: Systematic Analysis of Tumor Initiating Cells inHuman Colonic Adenocarcinoma Cell LineMaria Ausiliatrice Puglisi, Alessandro Sgambato, Marta Barba, Ezio Giorda, AlmaBoninsegna, Nathalie Saulnier, Luca Di Maurizio, Anna C. Piscaglia, Rita Carsetti, AchilleCittadini, Antonio Gasbarrini

BACKGROUND. The principal approaches for the isolation of cancer stem cells (CSC)fraction are through exclusion of the vital dye Hoechst 33342, which defines the so-calledside population (SP), or by the expression of cell surface markers, such as CD133. Ourresearch group have demonstrated the presence of a significant CD133+ cell fraction inhuman primitive and metastatic colon cancer (CC). Using the CACO2 cell line, we showedthat only CD133+ cells have a tumor-initiating potential In Vitro and In Vivo. However,other reports have not confirmed the role of CD133 as CSC-marker for CC. AIM. Thepresent study was aimed to identify new potential markers in order to increase the specificityin detecting CSC. METHODS. CD133+ and CD133- cells were isolated from CACO-2 cellline by FACS sorter and molecular profiling of both subpopulation was performed by themeans of microarray analysis. SP and non-SP fraction were isolated by FACS-sorter fromthe same cell line and characterized for the expression of CD133 and for tumorigenicpotential by In Vitro assays. Then, in the effort to identify a common “tumor stem cell”signature for CC, the most relevant transcripts resulting from gene expression profiling onCD133+ CACO2 cells was assessed by real-time PCR on SP-fraction. RESULTS: SP and non-SP fraction, isolated from CACO2 cell line, were characterized for the expression of stemcell marker CD133. Phenotypic analysis displayed that CD133 expression was higher in theSP fraction than no-SP fraction. Additionally, In Vitro assays showed that only SP cells wereable to generate clones on soft agar. Affymetrix molecular profiling of CD133+ versus CD133-caco-2 cells showed a significant overexpression of various transcripts involved in cellproliferation, invasion and stemness in CD133+ cell fraction. Real-time PCR analysis demon-strated that four of the most interesting transcripts identified in CD133+ CACO2 cells wereover-expressed in SP-fraction. CONCLUSION: Overall, SP and CD133 isolated from CACO-2 cell line, show analogue stem cell properties. These biological similarities are supportedby a similar gene expression profile, confirming that both approaches are valid methods forisolating CSC. In particular, four genes are highly expressed in both CD133 + cells and inSP cells. These genes are involved in regulating cell proliferation and metastatization, sothey may be excellent markers to increase the specificity of CD133 in identifying CSC in CC.

W1890

Dcamkl+1 and CD44 Positive Stem/Progenitor Cells are Significantly up-Regulated in Response to Autocrine Progastrin Expression, Associated WithSurprising Transformation of Hek-293 Cells, In Vitro and In VivoShubhashish Sarkar, Carla Kantara, Courtney W. Houchen, Shahid Umar, Pomila Singh

Background and Aims. We have reported that progastrin peptide (PG) exerts anti-apoptotic/proliferative effects on normal and cancerous epithelial cells, In Vitro and In Vivo, andincreases risk of colon carcinogenesis in transgenic mice overexpressing PG (Cancer 2004;Oncogene 2007; CanRes 2007; Oncogene 2008; JBC 2009). In Vivo studies with PG overex-pressing hGAS mice suggest an increase in DCAMKL+1/CD44 positive colonic crypt cells(Jin, et al, JCI 2009). It is, however, not known if PG directly induces an increase in stem/progenitor cells, In Vitro, and if this increase translates into transformation of the cells. Toaddress this novel possibility, we used PG responsive HEK-293 cells, which demonstratenegligible clonogenic/tumorigenic potential. Methods and Results. Mutant-hGAS-HEK293clones (HEK-PG) were generated and confirmed to overexpress full-length (9KDa) PG.HEK293 clones transfected with empty vector (HEK-C) served as controls. A significantincrease in proliferation of HEK-PG vs HEK-C cells was measured In Vitro. Cells, stained forthe putative stem cell marker (DCAMKL+1)/progenitor cell marker (CD44), were examined byfluorescense confocal microscopy and quantified by Western Blot analysis. Expression ofDCAMKL+1/CD44 increased 2-3 fold in HEK-PG clones vs HEK-C clones, suggesting forthe first time, that PG directly enhances the number of stem/progenitor cells. Significantincrease in DCAMKL+1/CD44 cells was confirmed in colonic crypts of Fabp-PG mice,overexpressing PG in colons. The most surprising finding was that overexpression of PG inHEK-PG clones imparted a clonogenic/tumorigenic/metastatic potential to the cells, In Vitroand In Vivo (using SCID mouse model). However, IEC-18-PG clones were not renderedclonogenic/tumorigenic, suggesting that overexpression of PG in the background of HEK-293 cells, represented the second hit required for transformation. IEC-18 cells are immortal-ized but apparently lack the first hit (initiating events) and remain non-transformed onPG over-expression. In associated studies, we have measured activation of two powerfultranscription factors or co-activzaators, NFκB and β-catenin, in response to PG, which mayexplain transformation of HEK-293-PG cells. Since a large % of GI cancers express autocrinePG, down-regulation of PG or its receptor (Annexin2), or the mediatory transcriptionalfactors (NFκB/β-catenin) should prove to be a powerful method for reducing stem cell/progenitor cell populations and attenuating tumorigenic/metastatic potential of PG-expressingcancer cells. Supported by NIH grants R01CA114264 and R01CA097959 and a Sealy Centergrant to P. Singh

S-760AGA Abstracts

W1891

Epithelial or Mesenchymal: Where to Draw the Line?Cristina Modak, Jennifer Pham, Jianyuan Chai

Background: Epithelial and mesenchymal cells represent two of the main differentiated celltypes in all vertebrates. However, their distinction is not always absolutely clear. Dozens ofmolecules have been claimed and used as markers for each cell type, while emerging evidencequestions their validity. The aims of this study are to compare the molecular phenotype ofepithelial cells with mesenchymal cells and to evaluate the molecular markers that arecommonly used to distinguish these two cell types. Methods: Six gut-derived human andrat cell lines, including epithelial cells and fibroblasts, were compared by real-time RT-PCRand immunofluorescence microscopy for the expressions of 11 epithelial markers and 11mesenchymal markers. The epithelial cells were also treated with TGFbeta-1 to inducemesenchymal phenotype. Results: All of the 22 “markers” tested were found in both epithelialand mesenchymal cells. Some epithelial markers, such as CLDN5, OCLN, DSG1 and TJP1,were even expressed at higher levels in fibroblasts than in epithelial cells. In comparison,mesenchymal markers showed more fidelity, but CDH2 and MMP9 were still significantlyhigher in epithelial cells than in mesenchymal cells. Furthermore, some “markers” alsoshowed individuality in response to TGFbeta-1 stimulation, for instance, TGFbeta-1 up-regulated epithelial markers such as CTNNB1 and CTNND1, but suppressed mesenchymalmarkers such as S100A4, FGF1 and FGF2. Conclusions: No gene expression is cell-typerestricted at least at the transcriptional level. Even though some of these “markers” areexpressed more in one cell type than in the other or appear in different cellular localization,none of them shows a unique pattern across species to make them universal markers.Nonetheless, some molecules appear to be better markers than others for a specific cell type.The information provided here is expected to serve as a reference for both basic scientistsand clinical researchers in the fields of epithelial-mesenchymal transition, molecular celltyping and cancer diagnosis.

W1892

Survival in Enteropathy Associated T-Cell Lymphoma: Type of Lymphoma isNot a Prognostic FactorLaura R. de Baaij, Jolanda M. van de Water, Wieke H. Verbeek, Chris J. Mulder

Introduction Although the WHO classification of Enteropathy Associated T-cell Lymphoma(EATL) only distinguishes between two types of EATL (i.e. Type I is associated with celiacdisease, Type II is not), in practice it appears that EATL Type I may present itself in twodifferent forms. For one, there is the EATL that develops after a period of refractory celiacdisease type II (EATL after RCDII). In the second form, EATL develops in patients withuncomplicated celiac disease, which is frequently diagnosed shortly before or at the sametime as EATL (EATL de novo). So far, data about survival rates in both groups are limited.Hitherto only one study has been performed to investigate the survival rate according tothe associated type of enteropathy, concluding that survival is significantly shorter in EATLafter RCDII (Malamut et al, Gut 2009) Aims We investigated the survival rates in EATLafter RCDII and in EATL de novo. Materials and Methods: We performed a retrospectivestudy of the medical files of 38 patients with EATL seen in our center between 2003 and2009. Patients with EATL Type II were excluded (n=2). Kaplan-Meier curves and the Log-rank test were used to compare survival between the two groups. Results: From a total of36 patients, 18 patients developed an EATL after RCD II (9M/9F) and 18 patients presentedwith EATL de novo (12M/6F). Mean age at diagnosis of EATL was 63 and 64 years respectively.In the group of 18 EATL after RCD II, 16 patients died. In the group of EATL de novo, 14of 18 patients died. In this latter group, 6 patients were treated with experimental stem celltransplantation, 5 of these patients died (after 6, 8.75, 11, 14.5, and 15 months). Themedian survival time in EATL after RCD II and EATL de novo is 5.0 months (95% CI:0.63-9.4) and 9.5 months (95% CI:1.8-17.2) respectively. Two-year survival in the EATLafter RCDII group is 18%, whereas in EATL de novo this is 25%. The difference in survivalbetween the two groups is not significant (p=0.351). Conclusions: In contrast to otherrecently published data, we have found no significant difference in survival rates betweenEATL after RCDII and EATL de novo. However, 6 of the patients with EATL de novo havebeen treated with experimental stem cell transplantation, this might have influenced theoverall survival rate in this group.

W1893

Skp2/Cks1 Overexpression and Cytoplasmic P27kip1 Immunoreactivity arePredictors of Poor Prognosis in Patients With Extrahepatic Bile Duct CancerHyunil Seo, Hongjoo Kim, Jung Ho Park, Dongil Park, Yong Kyun Cho, Chong Il Sohn,Yong Kyun Cho, Byung Ik Kim

Background and aims: S-phase kinase associated protein 2 (Skp2) is an F-box substrate-recognition subunit of the SCF ubiquitin-protein ligase complex, which regulates progressionof cell cycle by targeting p27kip1. Ubiquitination of p27kip1 also requires the presence ofcdc kinase subunit 1 (Cks1). Decreased levels of p27kip1 seem to be associated with highaggressiveness and poor prognosis in a variety of cancer. However, there have been fewreports concerning the expression status and the prognostic implication of Skp2 and Cks1in extrahepatic bile duct cancer. Methods: We have enrolled 76 patients who underwentcurative resection for histologically confirmed extrahepatic cholangiocarcinoma at KangbukSamsung Hospital from December 1994 to March 2008. Immunohistochemical staining forSkp2, Cks1, p27kip1 and Ki67 were performed using the relevant microarray tissue blocksof cholangiocarcinoma. The percentage score of staining positive tumor cells and the stainingintensity were multiplied to produce an immunoreactive score (IS) for each tumor specimen.Results: Immunohistochemical staining showed that Skp2, Cks1 and Ki67 proteins wereexpressed in the nuclei of cancer cells. The staining pattern of p27kip1 was somewhatheterogeneous, however, the pattern of nuclear staining was scarcely noticed, and most ofthem were expressed as diffuse cytoplasmic staining. The results of the immunohistochemicalstaining for Skp2, Cks1, Ki67 and p27kip1 were as follows. Skp2 intensity - 0, 32 (42.1%);1, 21 (27.6%); 2, 16 (21.1%); 3, 7 (9.2%): Cks1 intensity - 0, 9 (11.8%); 1, 24 (31.6%);