and IBD remain unclear. CMV infection may predispose some individuals to develop IBD,exacerbate existing IBD, and it is uncertain whether CMV is pathologic or reflects colonizationof rapidly dividing, dysplastic or inflamed tissue.While prior studies have suggested increasedcolectomy and mortality rates in IBD patients with CMV and improved remission rates withantiviral treatment, the impact of the severity of CMV infection on clinical outcome has notbeen examined. AIM: To evaluate whether severity of CMV colitis is associated with clinicaloutcomes in IBD patients. METHODS: A search of Cedars-Sinai Pathology database usingthe criteria “CMV” with “IBD,” “Crohn's disease” or “Ulcerative colitis” was performed.Patients were stratified into three histologic groups: 1) no evidence of CMV, 2) “lightinfection” defined by positive CMV immunohistochemistry without viral inclusion bodieson H&E stain, and 3) “heavy infection” defined by presence of viral inclusion bodies onH&E. Control subjects are IBD patients matched for modified HBI index. Charts werereviewed for modified HBI, mortality and colectomy rates, length of stay, and CMV quantitat-ive serum PCR. RESULTS: CMV infection was found in 27 of 603 patients (4.5%) andassociated with a higher colectomy rate compared to controls (48% vs. 30%, p=0.04).Colectomy rates were higher in patients with “heavy” (52%) compared to “light” (37%)infection. Antiviral treatment was associated with a 34% absolute reduction in the colectomyrate of patients with heavy, but not light, CMV infection. Surprisingly, peripheral CMV PCRlevels were lower in patients requiring colectomy (avg 63 copies, n=11) compared to thosewho avoided surgery (avg 198 copies, n=11). Among all antiviral treated patients, colectomyrate was higher in subjects with lower (avg 100 copies, n=6) compared to higher (avg 218copies, n=10) peripheral CMV levels. CONCLUSIONS: Heavy CMV infection is associatedwith higher colectomy rates. Importantly, peripheral CMV PCR levels do not correlate withhistological CMV infection or colectomy rate. Antiviral treatment may be more effective inIBD patients with heavy CMV infection. This suggest that in IBD patients with heavy CMVinfection, the virus may contribute more to the active inflammatory process and symptompresentation, than in those with lower CMV burden, where the bulk of symptoms are likelydue to underlying severe IBD itself.

1005

Clostridium difficile and Inflammatory Bowel Disease: Has the DiseaseChanged in Severity From 1998 to 2007?Ashwin N. Ananthakrishnan, Emily L. McGinley, Kia Saeian, David G. Binion

Introduction: Clostridium difficile (C difficile) has emerged as an important pathogen inpatients with inflammatory bowel disease (IBD) and is associated with increased morbidityand mortality. Although studies have documented an increase in prevalence, none hasexamined the temporal change in severity of C difficile infection (CDI) complicating IBD.Methods: Using data from the Nationwide Inpatient Sample, a national hospitalizationdischarge database in the United States, we identified all IBD-related hospitalizations duringthe years 1998, 2004, and 2007 using diagnosis codes. We examined hospitalizations witha co-existing code for C difficile. We performed two separate analyses: (1) Comparison ofthe actual outcomes (in-hospital mortality, colectomy rate) of C difficile-IBD across the threeyears; and (2) Comparison of these outcomes in the C difficile-IBD cohort relative to thenon-C difficile IBD controls in each corresponding year. Results: There were an estimated1,998, 4,800, and 6,902 IBD hospitalizations associated with CDI during 1998, 2004 and2007 respectively accounting for 1.4%, 2.3% and 2.9% of all IBD hospitalizations nationwide(p<0.001). Hospitalization during 2007 was associated with two-fold greater odds of CDI(Odds ratio (OR) 2.15, 95% confidence interval (CI) 1.92-2.41). In analysis restricted tothe C difficile-IBD population, hospitalization in 2004 (OR 1.38, 95% CI 0.83-2.28) or 2007(1.15, 95% CI 0.72-1.84) was not associated with higher mortality or need for colectomycompared to 1998. However, compared to the non-C difficile IBD controls during each ofthe corresponding time periods, there was an increase in the relative mortality risk associatedwith C difficile from 1998 (OR 2.38, 95% CI 1.52-3.72) to 2004 (4.00, 95% CI 3.03-5.29)and 2007 (OR 3.38, 95% CI 2.66-4.29). Similarly, there was relative increase in totalcolectomy risk from 1998 (OR 1.39, 95% CI 0.81-2.37) to 2004 (OR 1.62, 95% CI 1.15-2.29) and 2007 (OR 2.51, 95% CI 1.90-3.34). Conclusion: There has been a temporalincrease nationwide in CDI complicating IBD hospitalizations. The actual rate of adverseoutcomes in C difficile-IBD patients has remained relatively constant during between 1998and 2007. However, the risk of mortality relative to the non-C difficile-IBD cohort appearsto have increased between 1998 and 2004, and the relative risk of colectomy has alsocontinued to till 2007.

1006

Prevalence of Clostridium difficile Infection in Patients With InflammatoryBowel Disease is Significantly Related With Treatment ModalityChristian D. Stone, Neil A. Accortt, Simon H. Magowan

Background: Previous studies have shown that Clostridium difficile infection (CDI) is increas-ing among patients with inflammatory bowel disease (IBD). Infection is often community-acquired (CA) and is especially prevalent in ulcerative colitis. Possible contributing factorsto this pattern of risk include immunosuppressive (IS) medications and severity of disease.This study investigated whether risk of CDI in IBD is related to medication use and exploreddifferences between CA and hospital-acquired (HA) CDI. Methods: We pooled data fromtwo large administrative health claims databases to conduct a retrospective cohort analysisto determine the prevalence of C. difficile in IBD (ICD-9: 008.45). Inclusion criteria were≧2 IBD diagnoses (ICD-9: 555.x or 556.x) within 15 months of each other and a minimumof 21 months continuous enrollment. Concurrent medication use was documented andcategorized into four exposure groups: 1) no drug exposure [during enrollment]; 2) 5-ASA only; 3) 5-ASA and IS, (includes steroids and/or azathioprine, 6-mercaptopurine,methotrexate and cyclosporine); and 4) IS only. All subjects were followed forward in timeuntil the occurrence of the first C. difficile event, date of total colectomy, end of captureperiod (June 30, 2008) or end of enrollment. CA C. difficile was defined as any case confirmedoutside of the hospital, within 48 hours of hospital admission, or at least 60 days after ahospital discharge. All other cases were defined as HA. The log-rank test was used todetermine statistical significance. Results: 55,707 IBD subjects met the inclusion criteria.45.7% were male and 55.1% had ulcerative colitis. The majority of subjects (52.1%) received

S-145 AGA Abstracts

a 5-ASA in combination with another therapy while 18.1% received a 5-ASA only. Theprevalence of CDI in the 5-ASA only group was significantly lower than in the 5-ASA + ISgroup (p<0.01) and the IS only group (p<0.01) (Table). When limiting the analysis to CAC. difficile cases (65.9%), the prevalence was still significantly lower in the 5-ASA only group.Conclusions: Among IBD patients, the prevalence of CA CDI was 2-fold greater than HACDI. IBD patients taking IS drugs were at increased risk for CDI. The lowest risk was seenamong patients who received only 5-ASA compounds.

*p<0.05 using 5-ASA only as the comparator group

1007

Incidence of Inflammatory Bowel Disease Following the Walkerton Outbreakof Waterborne Bacterial GastroenteritisJohn K. Marshall, Marroon Thabane, Amit X. Garg, William F. Clark, Paul Moayyedi,Stephen M. Collins

Background and Aim: Acute gastroenteritis has been suggested to increase subsequent riskof IBD. InMay 2000, contamination ofmunicipal water by E. coli 0157:H7 andCampylobacterspecies caused a large outbreak of acute gastroenteritis (GE) in Walkerton, Ontario. TheWalkerton Health Study (WHS) was initiated to study related long-term health outcomesincluding inflammatory bowel disease (IBD). Previous genetic analyses of this well-definedcohort have associated post-infectious irritable bowel syndrome (PI-IBS) in this cohort withdefects in innate immunity and barrier function. Methods: WHS participants were askedabout diagnoses of IBD during standardized annual interviews from 2002 to 2008. Primarycare physicians were also contacted to identify subjects with IBD and provide supportingmedical records. An independent adjudication committee confirmed the timing and detailsof each diagnosis. Prevalence of IBD was assessed among subjects who resided in theWalkerton postal code during the outbreak. Incidence was assessed in a cohort of Walkertonresidents with no prior diagnosis of Crohn's disease (CD) or ulcerative colitis (UC). Results:Among 3047 eligible WHS participants (54.7% female; mean age 36.8±22.6), 16 cases ofIBD were confirmed on review of records (9 CD, 6 UC and 1 indeterminate colitis). Norecords were available to confirm or refute an additional 41 self-reported diagnoses. Underthe conservative assumption that non-participants did not have IBD, the estimated crudeprevalence of confirmed IBD in Walkerton was 525.1/105 (295.4/105 for CD vs. 196.9/105

for UC). Excluding diagnoses made before the outbreak or in the first year of follow-up,the annual incidence of IBD among 2031 subjects with GE during the outbreak was 49.2/105 vs. 14.5/105 among 983 subjects who remained well (incidence rate ratio 3.39; 95%CI 0.21, 13.81). The incidence of CD among subjects with GE was 35.2/105, while that forUC was 7.0/105. Estimates of prevalence and incidence of CD exceed published Canadianpopulation-based rates1 (233.7/105 and 13.4/105 respectively). Conclusion: The overallprevalence and incidence of CD in Walkerton is elevated. Although there is a trend to anincreased incidence of IBD among residents who suffered acute GE during the outbreak, itis not statistically significant. 1Bernstein CN et al, Am J Epidemiol 2006;101:1559-1568Funded by the Ontario Ministry of Health and Long-Term Care and the Crohn's and ColitisFoundation of Canada (CCFC)

1008

No Increase in JC Viremia, Lymphocyte Count, or Circulating CD34+Hematopoietic Progenitor Cells After Treatment With Vedolizumab, aHumanized Monoclonal Antibody to α4β7 IntegrinAsit Parikh, Tim Wyant, David B. Clifford, Joseph R. Berger, Brian G. Feagan, Irving Fox,Catherine Milch

Introduction: α4β7-Integrin on subsets of B- and T-lymphocytes mediates lymphocytetrafficking to the gastrointestinal tract. Vedolizumab is a selective humanized monoclonalantibody that blocks migration of α4β7-expressing lymphocytes to the gastrointestinal tractby antagonizing α4β7-MAdCAM interactions. Progressive multifocal leukoencephalopathy(PML), a rare but potentially fatal CNS infection caused by JC virus, is associated with thedual α4β1/α4β7 integrin antagonist natalizumab. Although the pathogenesis of PML is notestablished, elevations in circulating JC virus, total lymphocytes, and circulating CD34+hematopoietic progenitor cells have been hypothesized to contribute to PML associated withnatalizumab. Thus it is important to determine whether treatment with a selective α4β7integrin inhibitor causes these same biological changes. Methods: JC viremia was assessedretrospectively (assay sensitivity threshold 30 copies/ml) using frozen serum samples collectedat 2 month intervals from 1038 subjects enrolled in 9 vedolizumab clinical trials, includingongoing blinded phase 3 studies. Proportions of circulating CD34+ cells were assessedlongitudinally via FACS analysis of total leukocytes and lymphocytes from normal individualswho received a single IV dose of vedolizumab 300 mg. Results and Discussion: Approximately800 subjects have received 1-19 doses of vedolizumab for up to 2.5 years. JC virus testingshowed no association between vedolizumab exposure and viremia. Overall incidence of JCviremia was <1%with 5 unique patients testing positive (30-800 copies/ml) with no persistentviremia. All subjects with a positive result before or while receiving vedolizumab had asubsequent negative result. Furthermore, the occurrence of viremia did not increase withexposure duration. These findings suggest that JC viremia in subjects receiving vedolizumab,many in combination with corticosteroids or immunomodulators, may represent a stochasticevent not influenced by vedolizumab exposure. Absence of lymphocytosis was demonstrated

AG

AA

bst

ract

s

AG

AA

bst

ract

sby aggregate CBC analysis. In addition, no change was observed in relative frequency ofCD34+ cells in the peripheral circulation (baseline ~0.3%) either in the short term, at peakvedolizumab serum concentrations, or through Day 141, when drug has cleared in mostsubjects. Forced egress of leukocytes from bone marrow, CD34+ cells in particular, representsa potential means of JC virus reactivation. These data indicate important mechanistic distinc-tions between vedolizumab and natalizumab that may portend a lower risk of PML. To date,neither PML nor any other opportunistic infection has been reported for vedolizumab.

1009

Evaluation of the Incidence of Anti-JC Virus Antibodies in a Cohort ofNatalizumab-Treated PatientsAnthony M. Pepio, Simonetta Mocci, Julie Taylor, Lori Taylor, Ted Yednock, LeonidGorelik, Susan Goelz, Meena Subramanyam

Objective: Establish a validated assay to detect JC virus (JCV) antibodies in multiple sclerosis(MS) and Crohn's disease (CD) patients, and to evaluate JCV antibody status in patientswho later developed progressive multifocal leukoencephalopathy (PML). Background: JCvirus infection underlies the development of PML, a lytic infection of oligodendrocytes thatoccurs rarely in natalizumab-treated patients. Presence of JCV DNA in blood of patients hasnot been predictive of risk of developing PML. Methods: An ELISA for detecting anti-JCVantibodies in human serum and plasma was validated. The method was shown to be sensitive,specific and to have acceptable precision. An assay cut point was established using serafrom a broad panel of MS patients. Results: Evaluation of over 800 sera samples collectedfrom natalizumab-treated MS patients demonstrated the presence of distinct sero-positiveand sero-negative populations. Consistent with 2 recently published large studies in healthyvolunteers, 40-50% of these MS patients were classified as negative and 50-60% as positivefor anti-JCV antibodies. Antibody status of individual MS patients was generally stable overtime (~2 years) in both sero-positive and sero-negative groups. We evaluated serum samplesfrom 8 patients who developed PML while on natalizumab treatment. Samples from thesepatients had been collected 12-30 months prior to the diagnosis of PML and most werecollected prior to commencing natalizumab. In all 8 patients with PML, serum samples weresero-positive prior to onset of PML, suggesting pre-existing JCV infection in these patients.Sample analysis from additional populations of natalizumab-treated patients with MS andpatients with CD is ongoing. Conclusion: The validated assay provides a preliminary classifica-tion of patients as sero-negative or sero-positive with respect to anti-JCV antibodies. Patientswho developed PML had evidence of prior infection with JCV, as measured by the presenceof anti-JCV antibodies. Additional analyses are needed to determine whether the presence orabsence of anti-JCV antibodiesmay be useful in stratifying patients for risk of developing PML.

1010

Methadone, but Not Morphine, Inhibits Stimulated Chloride Currents in CLC-2 Transfected Cells in CultureJohn Cuppoletti, Danuta H. Malinowska, Kirti Tewari, Jayati Chakrabarti, Ryuji Ueno

Results from a recently completed clinical study suggest that concomitant use of methadone,but not other opiates such as morphine, ablates the effect of lubiprostone in treating thesymptoms of opioid-induced bowel dysfunction. The present study was undertaken to testthe hypothesis that therapeutic levels of methadone, but not morphine, prevent activationof ClC-2 chloride channels by lubiprostone. The effect of morphine and methadone wasalso studied on ClC-2 Cl- currents activated by forskolin and isobutylmethylxanthine (IBMX),which raise cAMP levels and activate PKA, an alternate mechanism of activating humanClC-2. Morphine and methadone effects were also examined on forskolin/IBMX-activatedCFTR Cl- currents. Methods: Whole cell patch clamp studies were performed to measureCl- currents in HEK293 cells transfected with human ClC-2 or with human CFTR. ClC-2-transfected HEK293 cells were treated with 100 nM lubiprostone, followed by treatmentwith 1 μM methadone and then 100 nM morphine or morphine and then methadone. Carewas taken to use concentrations of lubiprostone, morphine, and methadone that are withintherapeutic ranges used in humans. Effects of morphine and methadone were also examinedon ClC-2 Cl- currents and CFTR Cl- currents activated by 5 μM forskolin/20 μM/IBMX.Results: Lubiprostone significantly (p<0.001) stimulated Cl- currents at -140 mV in ClC-2 transfected HEK293 cells from -12.3 ± 8.3 to -123.0 ± 8.4 pA/pF (n=3). 100 nM morphinehad no effect, but methadone significantly inhibited Cl- currents to -62.0 ± 4.0 pA/pF(p<0.001). Methadone inhibited ClC-2 Cl- currents whether added before or after morphine.Forskolin/IBMX also significantly activated (p<0.001) ClC-2 Cl- currents from -20.3 ± 7.4to -104.6 ± 7.4 pA/pF (n=3), which was also inhibited by methadone irrespective of whetherit was added before (-58.5 ± 5.3 pA/pF; p<0.0005) or after (-67.9 ± 8.8 pA/pF; p<0.025)morphine. In contrast, neither methadone nor morphine affected CFTR Cl- currents whichwere activated from -23.8 ± 9.4 to -254.6 ± 21.5 pA/pF (n=4) with forskolin/IBMX. Conclu-sions: Methadone acts differently than morphine with respect to ClC-2. Methadone, butnot morphine, inhibits lubiprostone- and cAMP/PKA-stimulated Cl- currents in human ClC-2 transfected HEK293 cells. cAMP/PKA-stimulated CFTR Cl- currents were unaffected byeither morphine or methadone. These findings may explain the lack of lubiprostone effectwith concomitant methadone treatment in clinical trials and suggests that methadone inhibitsthe ClC-2 Cl- channel by a mechanism which is independent of lubiprostone, per se.Supported by Sucampo Pharmaceuticals, Inc.

1011

SLC26 Anion Transporters Involvement in Luminal Acid-Activated MurineDuodenal HCO3

- Secretion In VivoAnurag K. Singh, Brigitte Riederer, Anja Krabbenhoeft, Janina Bonhagen, Brigitte Rausch,Regina Engelhardt, Manoocher Soleimani, Ursula Seidler

Backround: HCO3- secretion protects the proximal duodenum against damage by gastric

acid. Duodenocytes express CFTR as well as at least three members of the SLC26 family(Slc26a3, 6 and 9) in the apical membrane. Aim and Methods: To establish the importance

S-146AGA Abstracts

of the different apical anion transporters during basal, acid and forskolin (FSK)- stimulatedduodenal HCO3

- secretion In Vivo, CFTR, Scl26a3, 6 and 9- deficient mice andWT littermateswere anesthetized, the proximal duodenum was luminally perfused with saline, 10-4 M FSK,or pH 2.2 for 5 min, and HCO3

- secretion was continuously titrated in the perfusate duringcontrolled systemic acid/base parameters. Gene expression was assessed by villus and cryptcell laser dissection and qPCR. Results: Basal duodenal HCO3

- secretion was slightly reducedin the Slc26a6-/-, and more strongly reduced in the Slc26a3-/-, Slc26a9, and CFTR-deficientmice compared toWT littermates. FSK-stimulated secretory responsewas normal in Slc26a6-/-

, Slc26a3-/-, Slc26a9-/- mice, as well as during luminal Cl- -free perfusion, but was virtuallyabolished in CFTR-deficient mice. Surprisingly, acid-activated HCO3

- secretion was unalteredin the Slc26a6-/- duodenum, but strongly reduced in Cl- -free luminal perfusion, Slc26a3-/-

and Slc26a9-/- duodenum, and abolished in the CFTR-deficient duodenum. 20μM of theCFTR(inh)-172 inhibited both acid- and FSK-stimulated HCO3

- secretory response by approx.50%. Laser dissection and immunohistochemistry revealed a villous-predominant expressionof Slc26a6, Slc26a3 and a crypt-predominant expression of CFTR and Slc26a9.Conclusions:Genetic deletion of intestinal Slc26 anion transporters reveals their differential involvementin basal, acid and FSK-stimulated HCO3

- secretion in murine duodenum. Most likely,different signalling in acid- vs. FSK-stimulated HCO3

- secretion explains this differentialinvolvement, rather than the differential crypt-villus expression pattern.

1012

Protein Tyrosine Phosphatase N2 Regulates TNFAlpha-Induced Signalling andCytokine Secretion in T84 Intestinal Epithelial CellsMichael Scharl, Declan F. McCole, Michael Fried, Gerhard Rogler

Background:We have previously shown that the Crohn's disease (CD) candidate gene, proteintyrosine phosphatase N2 (PTPN2), regulates interferon gamma (IFNγ)-induced signalling andepithelial barrier function in T84 intestinal epithelial cells (IEC). Studies in fibroblasts showedthat PTPN2 regulates tumor necrosis factor alpha (TNFα)-induced mitogen activated proteinkinase (MAPK) signalling. The aim of this study was to investigate whether PTPN2 isregulated by TNFα and if PTPN2 controls TNFα-induced signalling and effects in IEC.Methods: T84 IEC were used for all studies. Protein analysis was performed by Westernblotting, mRNA analysis by RT-PCR. PTPN2 knock-down was induced by siRNA andcytokine levels were measured by ELISA. Results: TNFα treatment (100 ng/ml) elevatedPTPN2 mRNA (6 h and 48 h; p<0.05; n=4) as well as nuclear (24 h; p<0.001; n=4)and cytoplasmic (72 h; p<0.001; n=4) protein levels. Immunofluorescence studies furtherindicated that TNFα causes cytoplasmic accumulation of PTPN2 by 72 h treatment. Inhibitionof nuclear factor κB (NF-κB) by the pharmacological inhibitor, BMS-345541 (p<0.001; n=3), completely prevented the TNFα-induced rise in PTPN2 protein (24 h; p<0.001; n=3).In contrast, though the MAPK/extracellular signal-regulated protein kinase (ERK) kinase(MEK) inhibitor, U0126, which also inhibits the activator protein-1 (AP-1) transcriptionfactor, blocked TNFα-induced ERK1/2 phosphorylation (p<0.01; n=3), it had no effect onPTPN2 protein (n=3). Knock-down of PTPN2 by siRNA revealed that the phosphatasedownregulates TNFα-induced ERK1/2 (24 h; p<0.01; n=3) and p38 (24 h; p<0.05; n=3)activity, without affecting c-Jun N-terminal kinase (JNK; n=3), inhibitor of κB (IκB; n=3),or NFκB (n=3) activity. On a functional level, loss of PTPN2 potentiated TNFα-inducedsecretion of interleukin (IL)-6 (p<0.01; n=5) and IL-8 (p<0.001; n=3) after 24 h. In TNFα(100 ng/ml; 24 h) and IFNγ (1000 U/ml; 24 h) co-treated cells, loss of PTPN2 enhancedthe expression of inducible nitric oxide synthase (iNOS) (p<0.05; n=3), and apoptosis asassessed by the amount of cleaved caspase-3 (p<0.05; n=3) and the number of fragmentednuclei in DAPI-stained cells. Conclusions: Our data demonstrate that TNFα induces PTPN2expression in T84 IEC via NFκB. Loss of PTPN2 promotes TNFα-induced MAPK signallingand expression of inflammatory mediators. We show that PTPN2 plays a key role in theregulation of TNFα-induced pro-inflammatory events in IEC, and thus PTPN2 activity mayplay an important role in the establishment of chronic inflammatory conditions in theintestine, such as CD. Supported by SNF and an educational grant by Essex Chemie, Switzer-land.

1013

The Apical Sodium-Dependent Bile Acid Transporter is Increased in HumanNecrotizing EnterocolitisSarah Mount Patrick, Jorn-Hendrik Weitkamp, Holly Dobrenen, Bohuslav Dvorak,Hernan Correa, Melissa D. Halpern

Background:Necrotizing enterocolitis (NEC) is themost common gastrointestinal emergencyof premature infants. We have previously shown that luminal and intra-enterocyte bile acids(BAs) are increased in experimental NEC and these increases precede ileal damage and theonset of inflammation. Further, the apical sodium-dependent bile acid transporter (ASBT),the ileal BA transporter responsible for active transport of BAs from the lumen into enterocytes,is upregulated in animals with experimental NEC. In contrast, the ileal bile acid bindingprotein (IBABP), a BA transporter within enterocytes, is decreased. These data stronglysuggest that BAs play a crucial role in the initiation of experimental NEC and that upregulationof ASBT may be a mechanism by which BAs accumulate within enterocytes. It is not known,however, if similar changes occur in human NEC. Objective: To examine ASBT and IBABPin human NEC. Methods: De-identified, archived human ileal sections from preterm infantswith NEC (n = 12), non-NEC diagnoses (NND; n = 5, congenital atresia, n = 3, spontaneousileal perforation, n = 2) or ileostomy take-downs (ITD; n = 10) were evaluated. Sectionswere stained with anti - human ASBT (P.A.Dawson, Wake Forest University) or IBABP(Sigma) antibodies, followed by biotinylated secondary antibody, Vectastain Elite ABCreagent, diaminobenzidine and counterstained with hematoxylin. Stained slides were exam-ined by a blinded observer; sections without discernible villous structure (n = 5) were notevaluated. Sections were scored based on the percent villi stained for ASBT or IBABP: 0,no staining; 1, ≤ 20% villi stained; 2, ≤ 40% villi stained; 3, ≤ 60% stained; 4, ≤ 80%stained; 5, > 80% stained. Median staining scores were calculated and the Mann-Whitneytest for non-parametric values was used for statistical analysis. Results: Ileal tissue frompre-term infants diagnosed with NEC had significantly higher ASBT staining scores thanthose with NND, ITD or ITD after NEC diagnosis. There was no significant difference in