Modification of Gastric Mucin Oligosaccharide Expression ... · Modification of Gastric Mucin...
Transcript of Modification of Gastric Mucin Oligosaccharide Expression ... · Modification of Gastric Mucin...
MM
Ca
*�
e
BmeaoMcmbOtmtomrppSaarwrflmfrfrCnsp
Mctemr
GASTROENTEROLOGY 2009;137:1061–1071
odification of Gastric Mucin Oligosaccharide Expression in Rhesusacaques After Infection With Helicobacter pylori
ARA L. COOKE,*,‡ HYUN JOO AN,§ JAEHAN KIM,� DON R. CANFIELD,¶ JAVIER TORRES,# CARLITO B. LEBRILLA,§
nd JAY V. SOLNICK*,‡,¶
Department of Internal Medicine and ‡Department of Medical Microbiology & Immunology, Center for Comparative Medicine, and §Department of Chemistry, andDepartment of Viticulture and Enology, University of California, Davis; ¶California National Primate Research Center, Davis, California; amd #Unidad de Investigación
n Enfermedades Infecciosas, Hospital de Pediatría, IMSS, México City, MexicosaMtegMl
hggptrmfMstisMbdotto
tsmstRa
L
BA
SIC–
ALI
MEN
TARY
TRA
CT
ACKGROUND & AIMS: Helicobacter pylori attaches toucin oligosaccharides that are expressed on host gastric
pithelium. We used the rhesus macaque model to char-cterize the effect of H pylori infection on gastric mucinligosaccharides during acute and chronic infection.ETHODS: Specific pathogen (H pylori)-free rhesus ma-
aques were inoculated with H pylori J166. Biopsy speci-ens of the gastric antrum were obtained 2 and 4 weeks
efore and 2, 8, and 24 weeks after infection with H pylori.-linked mucin oligosaccharides were released from gas-
ric biopsy samples by �-elimination and profiled byatrix-assisted laser desorption/ionization mass spec-
rometry. Similar studies were performed on gastric bi-psy samples from H pylori-infected and uninfected hu-ans. Formalin-fixed, paraffin-embedded sections of
hesus antrum biopsy samples were stained with H&E,eriodic acid-Schiff stain, and antibody to MUC5AC, theredominant mucin expressed in the stomach. RE-ULTS: H pylori-induced gastritis was accompanied byn acute and dramatic decrease in diversity and relativebundance of O-linked mucin oligosaccharides in thehesus stomach, which largely recovered during the 24-eek observation period. These variations in oligosaccha-
ide abundance detected by mass spectrometry were re-ected by changes in periodic acid-Schiff-positiveaterial and expression of MUC5AC over time. Relatively
ew differences were seen in gastric mucin oligosaccha-ide composition between H pylori-infected and unin-ected patients, which is consistent with the results inhesus macaques because infection occurs in childhood.ONCLUSIONS: Acute H pylori infection is accompa-ied by a dramatic but transient loss in mucin oligo-accharides that may promote colonization andersistence.
ucins are large glycoproteins that are the majorstructural component of the mucous gel layer that
overs the epithelial surface and serves complex func-ions, which include not only cytoprotection1 but alsopithelial growth, development, and protection againsticrobial pathogens.2 Mucins carry dense oligosaccha-
ide side chains connected by O-glycosidic linkages at
erine or threonine residues. In the healthy human stom-ch, the major mucin proteins are secreted, includingUC5AC, which is localized to surface epithelial cells of
he cardia, fundus, and antrum, and MUC6, which isxpressed in the neck cells of the fundus and in antrallands. The membrane-associated mucins, MUC1 andUC13, are also found in the stomach, although in
ower abundance.The gastric mucin population differs markedly in
ealth and disease. In gastric precancerous lesions and inastric cancer, there is altered expression of the usualastric mucins, MUC5AC and MUC6, as well as meta-lastic expression of the intestinal mucin, MUC2.3 Al-ered expression of gastric mucins has also been found inesponse to infection with Helicobacter pylori, which is the
ajor cause of peptic ulcer disease and an importantactor in the development of gastric cancer.4 Because the
UC5AC mucin displays Lewis b (Leb) and other fuco-ylated blood group antigens, which are thought to behe major receptors for H pylori in normal gastric tissue,5
t is perhaps surprising that H pylori infection hasometimes been associated with reduced expression of
UC5AC.6 – 8 However, contradictory findings have alsoeen reported in different study populations.9,10 Theseifferences may result in part from the fact that analysesf gastric mucin are limited by H pylori strain diversityhat exists in the human population and by the inabilityo study acute infection or even determine the durationf infection.
The nonhuman primate model offers the opportunityo study experimentally the effects of H pylori on expres-ion of mucin oligosaccharides. Rhesus monkeys express
ucins orthologous to human MUC5AC and MUC6 thathare similar density, size, glycoforms, oligomeric struc-ure, and tissue localization to those found in humans.11
hesus macaques also express ABO and Le blood groupntigens that serve as receptors for H pylori.11 Socially
Abbreviations used in this paper: CFU, colony-forming units; Leb,ewis b; MS, mass spectrometry; PAS, periodic acid-Schiff.
© 2009 by the AGA Institute0016-5085/09/$36.00
doi:10.1053/j.gastro.2009.04.014
hpadlmflaciucs
mtsiimdoeitsra
kit(cimSw
wpCaAwCCs
vtOwwt
bnMdmnfsnoaCtcod
ssaagasitthsprmptwpgVgm
BA
SIC–
ALIM
ENTA
RY
TRA
CT
1062 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
oused rhesus monkeys are naturally infected with Hylori,12 which supports the relevance of the model, but isn impediment to experimental infection. We thereforeerived specific-pathogen (H pylori)-free macaques by iso-
ating them at birth.12 Experimental H pylori infection inacaques is characterized by a T helper cell 1-type in-
ammatory response, with infiltration of mononuclearnd polymorphonuclear leukocytes that mimics thehronic active gastritis that is a hallmark of H pylorinfection in humans.13 Thus, the rhesus macaque can besed to study the role of gastric mucins in H pyloriolonization and persistence in an ecologically relevantystem.
We previously developed methods to analyze gastricucin oligosaccharides from rhesus monkeys using ma-
rix-assisted laser desorption/ionization (MALDI) masspectrometry (MS).14 Here we describe the use of MS andmmunohistochemistry to determine the effect of exper-mental challenge with H pylori on rhesus monkey gastric
ucins during acute and chronic infection. The resultsemonstrated a dramatic but transient decrease in mucinligosaccharides in the rhesus monkey stomach afterxperimental challenge with H pylori, which was reflectedn the amount of periodic acid-Schiff (PAS)-positive ma-erial and by changes in levels of the MUC5AC mucin. Wepeculate that H pylori modulates gastric mucin glycop-oteins during acute infection to promote colonizationnd persistent infection.
Materials and MethodsBacterial Strain and CultureH pylori J166 preferentially colonizes rhesus mon-
eys and contains both a functional cag pathogenicitysland and the s1m1 allele of the vacA cytotoxin.15 Bac-eria were cultivated on brucella agar or in brucella brothDifco Laboratories, Detroit, MI) containing 5% bovinealf serum (GibcoBRL, Gaithersburg, MD) and antibiot-cs (trimethoprim, 5 mg/L; vancomycin, 10 mg/L; poly-
yxin B, 2.5 IU/L; amphotericin B, 4 mg/L; all fromigma–Aldrich, St. Louis, MO), and incubated at 37°Cith 5% CO2.
AnimalsFour male rhesus macaques 3 to 4 years of age
ere confirmed to be uninfected with H pylori usingrotocols described previously12 and were housed at thealifornia National Primate Research Center, which isccredited by the Association for the Assessment andccreditation of Laboratory Animal Care. All proceduresere approved by the primate center Research Advisoryommittee and by the University of California, Davis,hancellor’s Animal Use and Care Administrative Advi-
ory Committee. w
Animal InoculationsA 50-mL liquid culture of H pylori J166 was culti-
ated overnight to an OD600 of approximately 0.4, cen-rifuged, and resuspended in fresh brucella broth to anD600 of 1.0. Each monkey was inoculated orogastricallyith 1 � 109 colony-forming units (CFU). The inoculumas examined by Gram’s stain, urease, and oxidase tests
o ensure a pure culture of H pylori.
Biopsies and Quantitative CulturesEach monkey underwent biopsy 2 and 4 weeks
efore H pylori inoculation and 2, 8, and 24 weeks posti-oculation using a pediatric gastroscope (Pentax FG-16X,ontvale, NJ) with a 1.8-mm biopsy forceps. The proce-
ure was performed under ketamine anesthesia (10g/kg of body weight intramuscularly) after an over-
ight fast. Seven antral biopsy specimens were collectedrom the stomach during each endoscopy. Two biopsyamples were placed in 250 �L brucella broth, homoge-ized with a sterile glass rod, and plated by serial dilutionn brucella agar supplemented with 5% bovine calf serumnd antibiotics. Plates were incubated at 37°C with 5%O2 for 5 to 6 days. H pylori colonies were identified in
he conventional manner by colony morphology, micros-opy, and biochemistry. CFU were counted, and CFU/gf tissue was determined for each monkey. The limit ofetection was approximately 102 CFU/g.
Histology and ImmunohistochemistryTwo formalin fixed, paraffin-embedded biopsy
amples from each monkey at each time point wereectioned (4 �m), deparaffinized, stained with H&E, andnalyzed blindly for inflammation (mononuclear cells)nd activity (polymorphonuclear leukocytes). Scores wereenerated as a composite of inflammation and activityccording to the modified Sydney system.16 For the PAStain, sections were treated with the PAS Stain Kit (Amer-can Master Tech Scientific, Inc, Lodi, CA) according tohe manufacturer’s directions. For immunohistochemis-ry, sections were deparaffinized and treated with 3%ydrogen peroxide for 30 minutes. After rehydration, thelides were subjected to antigen retrieval in citrate buffer,H 6.0, in a decloaker at 122°C. Slides were cooled tooom temperature, washed in PBS, and blocked for 30
inutes. Slides were incubated overnight at room tem-erature with monoclonal anti-human gastric mucin an-ibody (1:2000 dilution of clone 45M1; Sigma–Aldrich),ashed in PBS, and incubated for 1 hour at room tem-erature with biotinylated goat anti-mouse immuno-lobulin G diluted 1:1000 in PBS � ova albumin. Theectastain ABC Elite Kit (Vector Laboratories, Burlin-ame, CA) was used to detect all antibodies according toanufacturer’s instructions. Slides were counterstained
ith hematoxylin before dehydration and mounting.aGSpbsoawdmptsta
7dcatt1h4nsd
�uAwasctwtin
Hww
bAtwtba
mntrwaer5
aputPmsawcosbpawlssetstf
m4pbmvs
BA
SIC–
ALI
MEN
TARY
TRA
CT
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1063
Human Gastric Biopsy SamplesGastric biopsy samples were obtained from the
ntrum of patients undergoing upper endoscopy at theeneral Hospital of the Instituto Mexicano del Seguroocial, México City, Mexico. Two biopsy samples wererocessed for H pylori culture as for macaques, and 2iopsy samples were snap frozen in liquid nitrogen andtored at �80°C until processed for MS. Samples werebtained from H pylori uninfected patients (n � 8; meange � SD, 38.6 � 5.7 years) and from patients infectedith H pylori (n � 9; 37.9 � 5.4 years), which wasocumented by culture, histology, and serology usingethods previously described.17 All H pylori strains were
ositive for the cag pathogenicity island, which was de-ected by polymerase chain reaction as previously de-cribed.18 Informed consent was obtained from all par-icipants in accordance with standard procedurespproved by the local Institutional Review Board.
Release of O-Linked Oligosaccharides FromGastric Biopsy SamplesThree biopsy samples were placed in 250 �L of
0% ethanol, homogenized with a sterile glass rod, andialyzed in Slide-A-Lyzer Mini Dialysis Units (10,000 MWutoff; Pierce, Rockford, IL) against 2 L of nanopure H2Ot room temperature overnight. The material retained inhe dialysis unit was lyophilized, and 1–3 mg was addedo 500 �L of alkaline borohydride solution (mixture of.0 mol/L sodium borohydride and 0.1 mol/L sodiumydroxide, Sigma–Aldrich). The mixture was incubated at2°C for 12 hours in a water bath, and the reaction waseutralized by addition of 1.0 mol/L hydrochloric acidolution in an ice bath to destroy excess sodium borohy-ride.
Oligosaccharide Purification by PorousGraphitized Carbon-Solid Phase ExtractionO-linked oligosaccharides released by reductive
-elimination were purified by solid phase extractionsing a porous graphitized carbon cartridge (SPE-PGC;lltech Associates, Deerfield, IL). The cartridge wasashed with H2O followed by 0.05% (vol/vol) trifluoro-cetic acid in 80% acetonitrile (ACN)/H2O (vol/vol). Theolution of released oligosaccharide was loaded on theartridge and washed with nanopure water at 1 mL/mino remove salts and buffer. O-linked glycans were elutedith ACN diluted in H2O to 10%, 20%, or 40%, with 0.05%
rifluoroacetic acid. Each fraction was collected and driedn a centrivap apparatus. Fractions were reconstituted inanopure water prior to MS.
Mass Spectrometric AnalysisMass spectra were recorded on an external source
iResMALDI (IonSpec Corporation, Irvine, CA) equippedith a 7.0 Tesla magnet. The HiResMALDI was equipped
ith a pulsed Nd:YAG laser (355 nm). 2,5-Dihydroxy- tenzoic acid was used as a matrix (5 mg/100 mL in 50%CN in H2O) for positive and negative mode, respec-
ively. A saturated solution of NaCl in 50% ACN in H2Oas used as a cation dopant. The oligosaccharide solu-
ion (0.6 �L) was applied to the MALDI probe followedy matrix solution (0.6 �L). The sample was dried understream of air prior to MS.A HiResESI (IonSpec Corporation, Irvine, CA) instru-ent equipped with a 9.4 Tesla magnet and Picoview
ano-ESI source (New Objective, Woburn, MA) was usedo obtain more sensitive detection of acidic oligosaccha-ides. Sample solutions for the nanoelectrospray sourceere delivered using a standard 6-port switching valvend an Eksigent 1D pump (Eksigent Technologies, Liv-rmore, CA). Sample solutions were delivered with a flowate of 250 nL/min composed of 0.1% formic acid in0/50 water/ACN (vol/vol).
Statistical AnalysisBacterial density (CFU/g) was log transformed
nd analyzed by analysis of variance (ANOVA) with re-eated measures. Pair-wise comparisons were performedsing Student t test with the Bonferroni correction. His-ologic analysis of inflammation, and of staining withAS or antibody to MUC5AC, was analyzed using aixed model analysis of variance (SAS proc. Mixed, ver-
ion 9.2; SAS Inc, Cary, NC). Post hoc comparisonsmong time points were based on least squares meansith the Bonferroni correction. The number of oligosac-
haride compositions detected in rhesus monkey samplesver time was analyzed by ANOVA with repeated mea-ures. Post hoc comparisons among time points wereased on Dunnett test, and pair-wise comparisons wereerformed using Student t test. Hierarchical clusternalysis was performed using HCE 3.5 software (http://ww.cs.umd.edu/hcil/hce, University of Maryland, Col-
ege Park, MD) for the systematic comparison of masspectrometric data from H pylori-infected and uninfectedamples. For the calculation, the absolute intensity ofach peak was normalized by total ion intensity of spec-ra. Pearson correlation coefficient was used to calculateimilarity, and each cluster was updated and drawn usinghe average group linkage method. The 2-tailed P valueor statistical significance was .05.
ResultsQuantitative H pylori CulturesTwo antral biopsy specimens were used to deter-
ine quantitative bacterial load from each monkey 2 andweeks pre- and 2, 8, and 24 weeks postchallenge with H
ylori. All monkeys were culture negative for H pyloriefore challenge and were stably infected with approxi-ately 106 to 107 CFU/g of tissue throughout the obser-
ation period (Figure 1A). There were no statisticallyignificant differences in H pylori bacterial load among
he 3 postinfection time points.wfm
taan
Fpgmpcalp cale
BA
SIC–
ALIM
ENTA
RY
TRA
CT
1064 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
HistopathologyBiopsy sections from animals pre- and 2, 8, and 24
eeks postchallenge were stained with H&E and scoredor inflammation (mononuclear cells) and activity (poly-
igure 1. Quantitative H pylori culture and histopathology in antral gostinoculation. (A) H pylori bacterial load (mean � SE log10 CFU/g) waastritis score (inflammation and activity) for all animals at 2 weeks preinodified Sydney system. Gastritis was significantly increased after chalrechallenge histopathology demonstrated a significant increase in gahallenge. (C–F) Representative photomicrographs of H&E-stained secmarked inflammatory response 2 (D), 8 (E), and 24 (F) weeks after H
amina propria, displacing both foveolar and glandular epithelial profileersisted, its disruptive effect on gastric morphology was diminished. S
orphonuclear leukocytes) using the revised Sydney sys- t
em. Comparison of representative sections of gastricntrum before and after H pylori infection demonstrated
marked inflammatory response, consisting predomi-antly of plasma cells and lymphocytes, with some his-
biopsy samples of rhesus macaques obtained 2, 8, and 24 weeksble among the 3 postinfection time points. (B) Mean (�SE) compositetion and 2, 8, and 24 weeks postinoculation was quantitated using theas determined by a mixed model ANOVA. Pair-wise comparisons withat 2 (***P � .001), 8 (***P � .001), and 24 (**P � .005) weeks after
demonstrate that, compared with 2 weeks preinfection (C), there wasi challenge. At 2 weeks postinfection, the inflammation expanded the8- and particularly 24-weeks postinfection, although the inflammationbars, 200 �mol/L.
astrics staoculalengestritistionspylor
s. At
iocytes and neutrophils superficially, and a more in-
ttwwtmplpdAmlliwit
bcrtotcmsaHcBatgN(emaHmdlgaActtaos
fsficR
mMHstsgpdaTmtrsv3sar
Flp
BA
SIC–
ALI
MEN
TARY
TRA
CT
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1065
ensely lymphocytic infiltration, sometimes follicular, inhe deeper mucosa. Inflammation began as early as 2eeks postinfection and persisted throughout the 24-eek observation period, although it diminished over
ime (Figure 1C–F). At 2 weeks postinfection, the inflam-atory infiltrate can be seen to expand the lamina pro-
ria, displacing or destroying both foveolar and glandu-ar epithelial profiles. At 8 and particularly 24 weeksostinfection, although the inflammation persisted, itsisruptive effect on gastric morphology was diminished.composite Sydney score (Figure 1B) combining inflam-ation and activity was significantly increased after chal-
enge (P � .0005). Pair-wise comparisons with prechal-enge histopathology demonstrated a significant increasen gastritis at 2 (P � .005), 8 (P � .001), and 24 (P � .005)eeks after challenge. No evidence of gastric atrophy,
ntestinal metaplasia, or dysplasia was detected at anyime point.
MS Analysis of Mucin O-LinkedOligosaccharidesA mucin glycoprotein is composed of a peptide
ackbone that carries dense carbohydrate side chainsonnected by O-glycosidic linkages to serine or threonineesidues. The O-glycosidic linkage is alkali labile, andhus the carbohydrate side chains can be released asligosaccharide alditols by �-elimination with NaOH inhe presence of NaBH4. These mucin O-linked oligosac-harides exhibit vast diversity because of the variety ofonosaccharide compositions that make up the oligo-
accharide, as well as to differences in length, branching,nd linkages among the individual monosaccharides.owever, all O-linked oligosaccharides are built on 1 of 8
ore structures containing GalNAc in a nonreducing end.ecause the instrument used in this study has high massccuracy (�5 ppm with external calibration) and resolu-ion (�100,000 full width at half height), the exact oli-osaccharide composition with regard to hexose (Hex),-acetylhexosamine (HexNAc), sialic acid, and fucose
Fuc) is readily determined solely based on mass. Forxample, an oligosaccharide with quasimolecular ion at/z (mass/charge ratio) 1649.602 detected with a toler-
nce of 0.01 mass units must be 1 Fuc, 4 Hex, and 4exNAc. Accurate mass provides unambiguous assign-ent of glycan peaks, whereas the composition provides
irect evidence for whether the oligosaccharide is O-inked (derived from mucins) or N-linked (nonmucinouslycoproteins). O-linked oligosaccharides were fraction-ted with solvents of varying polarity (10%, 20%, 40%CN) to partition them into highly anionic and neutralomponents and then analyzed directly by MS in bothhe positive (cation) and negative (anion) modes. Becausehe 40% ACN fractions did not contain detectable glycansnd the 10% and 20% fractions generated similar profilesf predominantly small, neutral oligosaccharides and
ome acidic structures, the 10% ACN fractions were used tor all analyses (Supplementary Figure 1). Assignments ofelected oligosaccharide compositions were further con-rmed by tandem MS using infrared multiphoton disso-iation19 as described in Supplementary Methods andesults and shown in Supplementary Figure 2.
MS Profiles of O-Linked MucinOligosaccharides in Monkeys and HumansWe first examined oligosaccharide profiles in
onkeys and humans that were uninfected with H pylori.ucin core-type structures were identified that containedexNAc and Hex residues, either with or without fuco-
ylation (Supplementary Table 1), which were all consis-ent with O-linked, mucin-type oligosaccharides. Humanamples yielded a greater number and more diversity oflycans than did samples from monkeys (Figure 2, Sup-lementary Table 1). Most of the 67 oligosaccharidesetected in monkeys (86.6%) were found in at least 3 of 4nimals prior to infection (Figure 2A, Supplementaryable 1). Oligosaccharides detected in fewer than 3 ani-als typically had m/z greater than 2100, suggesting that
he sensitivity for detection of larger mucin oligosaccha-ides may be reduced. In contrast to monkeys, humanamples yielded a greater number of oligosaccharides (96s 67) and showed more biologic variability, with only6.5% of the 96 oligosaccharides detected in all 8 humanamples (Figure 2B, Supplementary Table 1). Humansnd monkeys shared approximately 39% of oligosaccha-ide compositions (Figure 3, Supplementary Table 1).
igure 2. Oligosaccharide masses (m/z) detected in 1 or more H py-ori-uninfected monkeys (A) or humans (B). Oligosaccharide masses arelotted against the percentage of individuals in which they were de-
ected.Ot1hLtaocehmm
msckcpi(siTs
sbpTdoccai
eotttl
4Fducwhct
BA
SIC–
ALIM
ENTA
RY
TRA
CT
1066 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
ligosaccharides were grouped according to whetherhey were common in humans and monkeys or unique to
species or another (Supplementary Table 1). Largeighly fucosylated structures with mixtures of Lea andeb epitopes were detected only in humans, althoughhey expressed fewer anionic oligosaccharides. The sialiccid N-glycolylneuraminic acid (NeuGc) was detectednly in monkeys (and NeuAc only in humans), which isonsistent with inactivation of the gene encoding thenzyme CMP-N-acetylneuraminic acid (CMP-Neu5Ac)ydroxylase, thought to have occurred approximately 2.1illion years ago before the origin of present-day hu-ans.20
Effects of H pylori Infection on MS Profileof O-Linked Mucin OligosaccharidesTo determine the effects of H pylori infection on
ucin oligosaccharides, we first examined whether oligo-accharide profiles were stable over time in rhesus ma-aques before H pylori challenge. Spectra from each mon-ey 2 and 4 weeks prior to H pylori inoculation wereompared for the presence or absence of oligosaccharideeaks and for the relative intensity of each peak normal-
zed to the total ion intensity of the respective spectrumSupplementary Figure 3). The glycan profiles were veryimilar for all animals between 2 and 4 weeks prior tonoculation, with correlation coefficients greater than 0.9.hese data indicate that oligosaccharide profiles are both
table and reproducible over time.We next compared the mean (�SD) number of oligo-
accharides detected in the spectra of the 4 monkeysefore challenge to that obtained 2, 8, and 24 weeksostchallenge with H pylori (Figure 4, Supplementaryables 2 and 3). After 2 weeks of infection, there was aramatic decline in mucin oligosaccharides, which couldnly be differentiated from background upon 5-fold con-entration of the samples. Number of oligosaccharideompositions recovered partially 8 weeks after infectionnd returned to preinoculation levels by 24 weeks ofnfection.
Comparison of the normalized peak intensities forach animal at each time point using a heat map dem-nstrated that, although the number of peaks returnedo preinoculation levels for all 4 monkeys, the peak in-ensities for 2 animals (animals C and D) differed fromhat detected prior to challenge (Figure 5). Thus, corre-ation coefficients of peak intensity between 4 weeks
™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™igure 3. Hierarchical cluster analysis of oligosaccharide masses (m/z)etected in H pylori uninfected monkeys and humans was performedsing HCE 3.5 software (http://www.cs.umd.edu/hcil/hce). Pearsonorrelation coefficient was used to calculate similarity, and each clusteras updated and drawn using the average group linkage method. Theeat map displays the percentage of samples in which each oligosac-haride was detected as a gradient from 100% (red) to 50% (black) and
o 0% (green).b(fioifppTcgioi
oospSdcwPtnIss
tmtvi(mqnmolwaTmsp
bwrtMfdtp
Fpaws
Ftadppia
BA
SIC–
ALI
MEN
TARY
TRA
CT
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1067
efore and 24 weeks after H pylori inoculation were highR � 0.97) for monkeys A and B but substantially loweror monkeys C (R � 0.61) and D (R � 0.66) (Figure 6),ndicating that, although a similar number and type ofligosaccharides were detected 24 weeks after infection,
n some cases the normalized absolute intensity differedrom the preinoculation levels. Comparison of glycanrofiles between patients infected or uninfected with Hylori showed relatively few differences (Supplementaryable 4), which is consistent with the monkey data be-ause infection occurs in childhood and is chronic. To-ether, these results demonstrate that H pylori infectionnduces an acute and dramatic decrease in gastric mucinligosaccharides, which largely recovers during chronic
nfection.
PAS StainWe next asked whether the changes in mucin-type
ligosaccharides detected by MS were supported by lossf carbohydrate content in gastric tissue sections. PAStains were performed on sections of gastric biopsy sam-les from monkeys before and after H pylori challenge.ections from a representative monkey demonstrated aramatic decrease in PAS staining 2 weeks after H pylorihallenge that partially reversed over the subsequent 24-eek observation period (Figure 7A–D). Quantitation ofAS-positive material was then performed on images ofhe stained sections using Image J software (http://rsbweb.ih.gov/ij/; National Institutes of Health, Bethesda, MD).
mages were cropped to retain only gastric mucosa andplit into red, green, and blue channels using the RGB
igure 4. Mean (�SD) number of oligosaccharide compositions de-ected in mass spectra from 4 monkeys before (�4, �2 weeks) andfter (2, 8, and 24 weeks) H pylori challenge. Mean number of peaksetected was significantly different among the 2-, 8-, and 24-week timeoints determined by ANOVA with repeated measures. Pair-wise com-arisons with prechallenge profiles demonstrated a significant decrease
n number of oligosaccharide compositions detected at 2 (**P � .005)nd 8 (***P � .001) weeks after challenge.
plit command. The green channel was used to calculate 1
he area of PAS-positive material because it providedaximal contrast between the PAS stain and the rest of
he mucosa. The area of PAS-positive material was di-ided by total area of the mucosal epithelium for eachmage, and the results at each time point were averagedFigure 7E). Percent PAS-positive material decreased
arkedly 2 weeks after H pylori challenge and subse-uently increased, although the preinfection level wasever recovered (Figure 7E). Statistical analysis by aixed model ANOVA demonstrated a significant effect
f time (P � .01). Pair-wise comparisons with prechal-enge biopsy samples showed that PAS-positive materialas significantly decreased at 2 weeks (P � .05) but nott 8 (P � .08) or 24 (P � 0.16) weeks after infection.hese results support the MS data and suggest thatucin oligosaccharide expression at the gastric epithelial
urface decreases during acute H pylori infection andartially recovers after 24 weeks.
ImmunohistochemistryImmunohistochemical analysis of paraffin-em-
edded gastric biopsy samples was used to determinehether the decrease in mucin-type oligosaccharides cor-
elated with changes in MUC5AC expression. Represen-ative images prior to H pylori challenge showed marked
UC5AC-positive staining, predominantly in surface/oveolar cells (Figure 8A). MUC5AC expression decreasedramatically after 2 weeks of infection (Figure 8B) andhen partially recovered over the 24-week observationeriod. Quantitation of MUC5AC staining using Image J
igure 5. Hierarchical cluster analysis of absolute oligosaccharideeak intensities detected in each monkey (A–D) at each time point (�4nd �2 and 2, 8, and 24 weeks) was performed using HCE 3.5 soft-are (http://www.cs.umd.edu/hcil/hce). The heat map displays the ab-olute ion intensity of each oligosaccharide peak as a gradient from
00% (red) to 50% (black) and to 0% (green).str
mclhtbcanteiuodmi
arsMadwpppsttdIsmmrfwa
BA
SIC–
ALIM
ENTA
RY
TRA
CT
1068 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
oftware as for PAS (Figure 8E) showed that, althoughhe overall trend was apparent, the time effect did noteach statistical significance (P � .18).
DiscussionAnalysis of the effects of H pylori on gastric
ucins using human clinical samples has resulted inonflicting reports, probably because these studies areimited by H pylori strain diversity that exists in theuman population and by the inability to determinehe duration of infection. For example, H pylori haseen associated with decreased MUC5AC and in-reased MUC6 expression in immunohistochemicalnalyses of gastric biopsy samples in some studies7 butot others.9,10 H pylori also inhibits total mucin syn-hesis in gastric epithelial cell lines,8 although mucinxpression in transformed cells is not necessarily phys-ologic. Here, we demonstrate for the first time in vivosing the rhesus macaque model that gastric mucinligosaccharide abundance and diversity dramaticallyecreases during the acute phase of infection and al-ost completely recovers to preinoculation levels dur-
ng chronic infection, despite persistent inflammation c
nd constant bacterial levels. Marked changes at highesolution were seen by MS and confirmed by PAStaining and immunohistochemical analysis of
UC5AC. These findings emphasize that a major ad-ptation occurs between H pylori and its mucosal nicheuring the first few weeks to months of infection,hich later reaches a state of equilibrium. We havereviously identified this to be a critical and dynamiceriod in studies of antigenic and phase variation in Hylori outer membrane proteins21 and host gene expres-ion studies of antimicrobial innate immune effec-ors.22 Mathematical models have also suggested thathe first few months of H pylori colonization are aynamic period in relation to the host response.23
ndividual differences in bacterial or host adaptation,uch as changes in expression of Lewis antigens,24,25
ay underlie the different glycan patterns among theonkeys (Figure 6). Because the human samples likely
epresent chronic colonization that has been presentor decades, it is not surprising that few differencesere found between H pylori-infected and uninfecteddults. The greater abundance and complexity of mu-
Figure 6. Correlation betweenmass spectra generated fromeach monkey (A–D) 4 weekspre- and 24 weeks postinocula-tion with H pylori. The absoluteintensity of each oligosaccharidepeak was first normalized by to-tal ion intensity of the spectrum,and then Pearson r was used tocalculate correlation betweenthe profiles generated for eachmonkey.
in oligosaccharides in humans compared with ma-
crf
as
pobssd
F2wP8
Fwsa
BA
SIC–
ALI
MEN
TARY
TRA
CT
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1069
aques might be a species difference but may alsoesult simply from the closed population of macaquesrom which the samples were drawn.
Previous work in the rhesus macaque demonstratedtransient decrease in mucosal Leb and increase in
ialylation in rhesus macaques 1 to 4 weeks after H
igure 7. Representative photomicrographs of PAS-stained sections of g4 weeks (D) postinoculation with H pylori. Percent of total area of gastricith Image J software (http://rsbweb.nih.gov/ij/) (E). Statistical analysis bair-wise comparisons with prechallenge biopsy samples showed that PA(P � .08) or 24 (P � .16) weeks after infection. Scale bar, 200 �m.
igure 8. Representative photomicrographs of MUC5AC-stained seceeks (C), and 24 weeks (D) postinoculation with H pylori. Percent of totaining) was quantitated with Image J software (http://rsbweb.nih.gov/i
lthough the overall trend was apparent, the time effect did not reach statistylori challenge.26 Similarly, we found a transient lossf MUC5AC, which displays Leb and other fucosylatedlood group antigens.5 However, sialylated glycans werecarce in the pool of total mucin oligosaccharides, repre-enting less than 3% of the total oligosaccharide abun-ance detected by MS, and no changes were found after
antrum from 2 weeks preinoculation (A) and 2 weeks (B), 8 weeks (C), andsa that stained PAS positive (bright magenta staining) was quantitatedixed model ANOVA demonstrated a significant effect of time (P � .01).sitive material was significantly decreased at 2 weeks (P � .05) but not at
of gastric antrum from 2 weeks preinoculation (A) and 2 weeks (B), 8ea of gastric mucosa that stained with anti-MUC5AC antibody (brownQuantitation of MUC5AC staining using Image J software showed that,
astricmucoy a mS-po
tionstal ar
j/) (E).
ical significance (P � .18). Scale bars, 200 �m.iult
stsgltdibpdtstasttwnii
oadbctemmtiipcBpacb
aG1
1
1
1
1
1
1
1
1
1
1
BA
SIC–
ALIM
ENTA
RY
TRA
CT
1070 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
nfection. Thus, it appears that sialylated glycans arencommon in the gastric mucosa, and changes in sialy-
ation may be detectable only with immunohistochemis-ry or other highly sensitive methods.
The mechanisms by which H pylori might alter expres-ion of mucin oligosaccharides are unknown. Choleraoxin and Entamoeba histolytica trophozoites enhance theecretion of preformed and newly synthesized mucinlycoproteins, which may deplete the protective mucusayer and facilitate pathogenesis.27,28 On the other hand,he available evidence suggests that H pylori actuallyown-regulates mucin exocytosis.29 Transcriptional changes
n mucin expression occur commonly in response toacterial products such as lipopolysaccharides, neutro-hil elastase, interleukin-1�, and other inflammatory me-iators, but, in contrast to our findings, the expression isypically induced.30 Reports of H pylori protease, glyco-ulfatase, and neuraminidase activities that degrade gas-ric mucins31,32 are controversial. Finally, H pylori maylter expression of genes involved in glycan synthesis,uch as induction of a GlcNAc transferase essential forhe biosynthesis of Lewis antigens.33 Because the reduc-ion in PAS was more striking than that for MUC5AC,hich was detected with an antibody thought to recog-ize the peptide backbone,34 we speculate that H pylori
nduces changes in mucin glycosylation during acutenfection.
The dramatic changes in mucin oligosaccharides webserved during acute H pylori infection could representpathogenic mechanism for bacterial persistence, a hostefense, or both. On the one hand, depletion of mucinsearing fucosylated blood group antigens and other gly-an receptors might enhance bacterial access to the epi-helial surface and promote chronic infection. Once thepithelial layer is colonized, there may no longer be anyicrobial benefit to mucin depletion, and, in fact, theucin layer may then be protective. On the other hand,
he decreased oligosaccharide content during acutenfection could be a host defense, much like MUC1 thats thought to act as a “releasable decoy ligand” for Hylori that, upon bacterial adherence, is cleaved by hostell proteases and released from the host cell surface.35
ecause the mucin oligosaccharides recover almost toreinoculation levels, despite persistent inflammationnd constant bacterial load, we speculate that thehanges seen during acute infection are likely critical foracterial colonization and persistence.
Supplementary Data
Note: To access the supplementary materialccompanying this article, visit the online version ofastroenterology at www.gastrojournal.org, and at doi:
0.1053/j.gastro.2009.04.014.References
1. Laine L, Takeuchi K, Tarnawski A. Gastric mucosal defense andcytoprotection: bench to bedside. Gastroenterology 2008;135:41–60.
2. Lievin–Le Moal V, Servin AL. The front line of enteric host defenseagainst unwelcome intrusion of harmful microorganisms: mucins,antimicrobial peptides, and microbiota. Clin Microbiol Rev 2006;19:315–337.
3. Hollingsworth MA, Swanson BJ. Mucins in cancer: protection andcontrol of the cell surface. Nat Rev Cancer 2004;4:45–60.
4. Kusters JG, van Vliet AH, Kuipers EJ. Pathogenesis of Helicobac-ter pylori infection. Clin Microbiol Rev 2006;19:449–490.
5. Van de Bovenkamp JH, Mahdavi J, Korteland–Van Male, et al. TheMUC5AC glycoprotein is the primary receptor for Helicobacterpylori in the human stomach. Helicobacter 2003;8:521–532.
6. Beil W, Enss ML, Muller S, et al. Role of vacA and cagA inHelicobacter pylori inhibition of mucin synthesis in gastric mu-cous cells. J Clin Microbiol 2000;38:2215–2218.
7. Byrd JC, Yan P, Sternberg L, et al. Aberrant expression of gland-type gastric mucin in the surface epithelium of Helicobacterpylori-infected patients. Gastroenterology 1997;113:455–464.
8. Byrd JC, Yunker CK, Xu QS, et al. Inhibition of gastric mucinsynthesis by Helicobacter pylori. Gastroenterology 2000;118:1072–1079.
9. Kang HM, Kim N, Park YS, et al. Effects of Helicobacter pyloriInfection on gastric mucin expression. J Clin Gastroenterol 2008;42:29–35.
0. Marques T, David L, Reis C, et al. Topographic expression ofMUC5AC and MUC6 in the gastric mucosa infected by Helicobac-ter pylori and in associated diseases. Pathol Res Pract 2005;201:665–672.
1. Linden S, Boren T, Dubois A, et al. Rhesus monkey gastricmucins: oligomeric structure, glycoforms and Helicobacter pyloribinding. Biochem J 2004;379:765–775.
2. Solnick JV, Canfield DR, Yang S, et al. Rhesus monkey (Macacamulatta) model of Helicobacter pylori: noninvasive detection andderivation of specific-pathogen-free monkeys. Lab Anim Sci1999;49:197–201.
3. Mattapallil JJ, Dandekar S, Canfield DR, et al. A predominant Th1type of immune response is induced early during acute Helico-bacter pylori infection in rhesus macaques. Gastroenterology2000;118:307–315.
4. Cooke CL, An HJ, Kim J, et al. Method for profiling mucin oligo-saccharides from gastric biopsy of rhesus monkeys with andwithout Helicobacter pylori Infection. Anal Chem 2007;79:8090–8097.
5. Solnick JV, Hansen LM, Canfield DR, et al. Determination of theinfectious dose of Helicobacter pylori during primary and second-ary infection in rhesus monkeys (Macaca mulatta). Infect Immun2001;69:6887–6892.
6. Dixon MF, Genta RM, Yardley JH, et al. Classification and gradingof gastritis. The updated Sydney System. International Workshopon the Histopathology of Gastritis, Houston 1994. Am J SurgPathol 1996;20:1161–1181.
7. Camorlinga–Ponce M, Flores–Luna L, Lazcano–Ponce E, et al.Age and severity of mucosal lesions influence the performance ofserologic markers in Helicobacter pylori-associated gastroduode-nal pathologies. Cancer Epidemiol Biomarkers Prev 2008;17:2498–2504.
8. Tummuru MK, Cover TL, Blaser MJ. Cloning and expression of ahigh-molecular-mass major antigen of Helicobacter pylori: evi-dence of linkage to cytotoxin production. Infect Immun 1993;61:1799–1809.
9. Lancaster KS, An HJ, Li B, et al. Interrogation of N-Linked oligo-saccharides using infrared multiphoton dissociation in FT-ICR
mass spectrometry. Anal Chem 2006;78:4990–4997.2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
R
fC7
C
F
C– YTR
ACT
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1071
0. Chou HH, Hayakawa T, Diaz S, et al. Inactivation of CMP-N-acetylneuraminic acid hydroxylase occurred prior to brain expan-sion during human evolution. Proc Natl Acad Sci U S A 2002;99:11736–11741.
1. Solnick JV, Hansen LM, Salama NR, et al. Modification of Heli-cobacter pylori outer membrane protein expression during exper-imental infection of rhesus macaques. Proc Natl Acad Sci U S A2004;101:2106–2111.
2. Hornsby MJ, Huff JL, Kays RJ, et al. Helicobacter pylori inducesan antimicrobial response in rhesus macaques in a cag patho-genicity island-dependent manner. Gastroenterology 2008;134:1049–1057.
3. Blaser MJ, Kirschner D. Dynamics of Helicobacter pylori coloni-zation in relation to the host response. Proc Natl Acad Sci U S A1999;96:8359–8364.
4. Wirth HP, Yang M, Sanabria-Valentin E, et al. Host Lewis pheno-type-dependent Helicobacter pylori Lewis antigen expression inrhesus monkeys. FASEB J 2006;20:1534–1536.
5. Nilsson C, Skoglund A, Moran AP, et al. Lipopolysaccharide di-versity evolving in Helicobacter pylori communities through ge-netic modifications in fucosyltransferases. PLoS ONE 2008;3:e3811.
6. Linden S, Mahdavi J, Semino-Mora C, et al. Role of ABO secretorstatus in mucosal innate immunity and H pylori infection. PLoSPathog 2008;4:e2.
7. Chadee K, Keller K, Forstner J, et al. Mucin and nonmucinsecretagogue activity of Entamoeba histolytica and cholera toxinin rat colon. Gastroenterology 1991;100:986–997.
8. Leitch GJ. Cholera enterotoxin-induced mucus secretion and in-crease in the mucus blanket of the rabbit ileum in vivo. InfectImmun 1988;56:2871–2875.
9. Micots I, Augeron C, Laboisse CL, et al. Mucin exocytosis: amajor target for Helicobacter pylori. J Clin Pathol 1993;46:241–
245. a0. Li JD, Dohrman AF, Gallup M, et al. Transcriptional activation ofmucin by Pseudomonas aeruginosa lipopolysaccharide in thepathogenesis of cystic fibrosis lung disease. Proc Natl Acad SciU S A 1997;94:967–972.
1. Sidebotham RL, Batten JJ, Karim QN, et al. Breakdown of gastricmucus in presence of Helicobacter pylori. J Clin Pathol 1991;44:52–57.
2. Suerbaum S, Friedrich S. Helicobacter pylori does not have a hapmucinase gene that is quasi-identical to the Vibrio cholerae hapgene. Mol Microbiol 1996;20:1113–1114.
3. Marcos NT, Magalhaes A, Ferreira B, et al. Helicobacter pyloriinduces �3GnT5 in human gastric cell lines, modulating expres-sion of the SabA ligand sialyl-Lewis x. J Clin Invest 2008;118:2325–2336.
4. Lidell ME, Bara J, Hansson GC. Mapping of the 45M1 epitope tothe C-terminal cysteine-rich part of the human MUC5AC mucin.FEBS J 2008;275:481–489.
5. McGuckin MA, Every AL, Skene CD, et al. Muc1 mucin limits bothHelicobacter pylori colonization of the murine gastric mucosa andassociated gastritis. Gastroenterology 2007;133:1210–1218.
Received January 13, 2009. Accepted April 9, 2009.
eprint requestsAddress requests for reprints to: Jay V. Solnick, MD, PhD, Center
or Comparative Medicine, University of California-Davis, Davis,alifornia 95616. e-mail: [email protected]; fax: (530)52-7914.
onflicts of interestThe authors disclose no conflicts.
undingSupported by Public Health Service Grants AI42081, GM490077,
nd NRSA AI60555.
BA
SIA
LIM
ENTA
R
smelTwfl1nlumecdwrno
sttoa((mppFHatrtks1
1071.e1 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
Supplementary Methods
Structural Elucidation of OligosaccharidesUsing Infrared Multiphoton DissociationTandem mass spectrometry was performed using
tructural elucidation of oligosaccharides using infraredultiphoton dissociation (IRMPD) to determine the gen-
ral structures of selected oligosaccharides, which al-owed for complete fragmentation of the ion of interest.he ion of interest was readily selected in the analyzerith the use of an arbitrary-wave form generator and a
requency synthesizer. A continuous wave Parallax CO2
aser (Waltham, MA) with 20-W maximum power and0.6-�m wavelength was installed at the rear of the mag-et and was used to provide the photons for IRMPD. The
aser beam diameter was 6 mm as specified by the man-facturer. The laser beam was expanded to �12 mm byeans of a 2� beam expander (Synrad, Mukilteo, WA) to
nsure complete irradiation of the ion cloud through theourse of the experiment. The laser was aligned andirected to the center of the ICR cell through a BaF2
indow (Bicron Corporation, Newbury, OH). Photon ir-adiation time was optimized to produce the greatestumber and abundance of fragment ions. The laser was
perated at an output of approximately 13 W. mSupplementary Results
Structural Elucidation of OligosaccharidesUsing Infrared Multiphoton DissociationTo confirm glycan composition and to obtain
tructural information, selected ions were subjected toandem mass spectrometry (MS) using infrared mul-iphoton dissociation (IRMPD) structural elucidation ofligosaccharides. For example, IRMPD analysis of the iont m/z 1430.528 ([M�Na]�), corresponding to 3 hexoseHex), 3 N-acetylhexosamine (HexNAc), and 2 fucoseFuc), yielded loss of 2 consecutive Fuc from the quasi-
olecular ion (m/z 1430¡1284¡1138), indicating theresence of 2 Fuc residues with nonreducing terminalositions and at least 1 branched point (Supplementaryigure 2). Similarly, the loss of Hex (m/z 1268) andexNAc (m/z 1227) indicated that these residues were
lso present as nonreducing termini. There was no fur-her Hex loss from m/z 1268, suggesting that the 2 Hexesidues were internal. The ion at m/z 611 correspondedo 1 Hex, 1 HexNAc, and HexNAc-ol, which identified thenown trisaccharide core structure for mucin-type oligo-accharides. Thus, the primary sequence of the ion m/z430 based on the tandem MS was determined (Supple-
entary Figure 2, inset).generated from each animal at each time point.
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1071.e2
Supplementary Figure 1. Mass spectra
S1
1071.e3 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
upplementary Figure 2. Structural elucidation of oligosaccharides using infrared multiphoton dissociation (IRMPD) analysis of the ion at m/z
430.528 ([M�Na]�). The primary sequence of the ion m/z 1430 based on the tandem MS is displayed in the inset.S
a
b
S
c
a
b
cOligosaccharides with higher detection frequency in monkeys.
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1071.e4
upplementary Table 1. Comparison of OligosaccharideMass, Composition, and Frequencyin Antral Gastric Biopsies ofHumans and Monkeys WithoutH pylori Infectiona,b,c
m/z [M�Na]
OS composition Frequency (%)
Hex HexNAc Fucose Human Monkey
Group I 611.228 1 2 0 100 100773.280 2 2 0 100 100814.307 1 3 0 100 100919.338 2 2 1 100 100960.365 1 3 1 100 100976.360 2 3 0 100 100
1065.396 2 2 2 100 1001122.418 2 3 1 100 1001138.413 3 3 0 100 1001179.439 2 4 0 100 1001268.476 2 3 2 100 1001284.470 3 3 1 100 1001325.497 2 4 1 100 1001341.492 3 4 0 100 1001430.528 3 3 2 100 1001487.550 3 4 1 100 1001544.571 3 5 0 100 1001633.608 3 4 2 100 1001690.629 3 5 1 100 1001795.661 4 4 2 100 1001852.682 4 5 1 100 1001941.718 4 4 3 100 1001998.740 4 5 2 100 100
Group II 757.285 1 2 1 100 881227.449 3 2 2 100 881471.555 2 4 2 100 881576.586 3 3 3 100 881836.687 3 5 2 100 881779.666 3 4 3 100 752144.798 4 5 3 100 632201.819 4 6 2 100 381747.651 3 6 0 100 02055.761 4 6 1 100 02087.776 4 4 4 100 01982.745 3 5 3 88 501373.507 3 2 3 88 01617.613 2 4 3 88 02290.856 4 5 4 88 02452.909 5 5 4 75 501211.454 2 2 3 75 02233.834 4 4 5 75 01414.533 2 3 3 63 02598.966 5 5 5 63 02509.930 5 6 3 50 38
554.206 1 1 1 50 01925.724 3 4 4 50 02347.877 4 6 3 50 02655.988 5 6 4 38 131763.671 2 4 4 38 01900.692 5 3 3 38 01366.524 1 5 1 25 01535.560 4 2 3 25 01731.656 2 6 1 25 01884.697 4 3 4 25 02046.750 5 3 4 25 02745.024 5 5 6 25 0
408.148 1 1 0 13 0
upplementary Table 1. (Continued)
m/z [M�Na]
OS composition Frequency (%)
Hex HexNAc Fucose Human Monkey
1097.386 4 2 0 13 01243.444 4 2 1 13 01300.465 4 3 0 13 01389.502 4 2 2 13 01512.582 1 5 2 13 01608.576 5 3 1 13 01754.634 5 3 2 13 01868.702 3 3 5 13 02112.782 4 7 0 13 02265.824 6 4 3 13 02395.887 5 4 5 13 02427.877 7 4 3 13 02557.940 6 4 5 13 02573.935 7 4 4 13 02630.956 7 5 3 13 02964.099 6 6 5 13 0
Group III 1081.391 3 2 1 88 1001649.603 4 4 1 88 1001592.581 4 3 2 50 1001017.386 1 4 0 38 1001446.523 4 3 1 25 1001503.545 4 4 0 25 1001738.639 4 3 3 25 1001811.656 5 4 1 13 1001957.713 5 4 2 13 1001382.519 2 5 0 0 100935.333 3 2 0 38 88
2306.851 5 5 3 75 882103.771 5 4 3 13 882014.735 5 5 1 50 832160.793 5 5 2 50 831528.577 2 5 1 25 751893.709 3 6 1 0 751706.624 4 5 0 38 632363.872 5 6 2 25 632249.829 5 4 4 13 632322.846 6 5 2 13 502468.904 6 5 3 13 50449.175 0 2 0 13 38
2614.961 6 5 4 13 382671.983 6 6 3 0 252818.041 6 6 4 0 252525.925 6 6 2 0 132566.952 5 7 2 0 13
Oligosaccharides detected in 100% of monkeys and humans.Oligosaccharides with higher detection frequency in humans.
SC
1071.e5 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
upplementary Figure 3. Reproducibility of glycan profiles generated from the same monkey at 2 time points prior to inoculation with H pylori.
orrelation was calculated from absolute peak intensities using Pearson r.S
a
b
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1071.e6
upplementary Table 2. Neutral Oligosaccharide Mass, Composition, and Frequency of Detection in Antral Gastric BiopsySamples From 4 Rhesus Monkeys Pre- and Post-H pylori Infectiona,b
Group m/z [M�Na]
OS composition
Frequency (%)
Pre Post
Hex HexNAc Fucose �4 wk �2 wk 2 wk 8 wk 24 wk
Group I 408.148 1 1 0 100 100 0 0 100449.175 0 2 0 100 100 0 0 100773.280 2 2 0 100 100 0 0 75
1341.492 3 4 0 100 100 0 0 751446.523 4 3 1 100 100 0 0 751503.545 4 4 0 100 100 0 0 751544.571 3 5 0 100 100 0 0 751592.581 4 3 2 100 100 0 0 751690.629 3 5 1 100 100 0 0 751852.682 4 5 1 100 100 0 0 751998.740 4 5 2 100 100 0 0 751957.713 5 4 2 100 100 0 0 502103.771 5 4 3 100 100 0 0 502160.793 5 5 2 100 100 0 0 502468.904 6 5 3 100 50 0 0 251528.577 2 5 1 75 75 0 0 751836.687 3 5 2 75 100 0 0 751893.709 3 6 1 75 75 0 0 751811.656 5 4 1 75 100 0 0 502014.735 5 5 1 75 100 0 0 50
935.333 3 2 0 75 75 0 0 502144.798 4 5 3 75 75 0 0 501982.745 3 5 3 75 50 0 0 502818.041 6 6 4 75 50 0 0 502322.846 6 5 2 75 50 0 0 252363.872 5 6 2 75 50 0 0 252509.930 5 6 3 75 25 0 0 252614.961 6 5 4 75 25 0 0 251706.624 4 5 0 50 50 0 0 252201.819 4 6 2 50 50 0 0 252525.925 6 6 2 50 25 0 0 252671.983 6 6 3 50 25 0 0 252655.988 5 6 4 50 0 0 0 252834.036 7 6 3 25 0 0 0 252980.094 7 6 4 25 0 0 0 252566.952 5 7 2 25 0 0 0 02713.010 5 7 3 25 0 0 0 01471.555 2 4 2 75 75 0 25 751325.497 2 4 1 100 100 25 0 751649.603 4 4 1 100 100 25 0 751738.639 4 3 3 100 100 25 0 501941.719 4 4 3 100 100 25 0 502306.851 5 5 3 100 100 25 0 501779.666 3 4 3 75 75 25 0 502249.829 5 4 4 75 75 25 0 502452.909 5 5 4 75 50 25 0 501017.386 1 4 0 100 100 25 25 75
Group II 1081.391 3 2 1 100 100 0 50 75757.286 1 2 1 100 75 0 50 75960.365 1 3 1 50 100 0 50 75611.228 1 2 0 100 100 0 75 75
1576.586 3 3 3 75 100 25 50 50919.338 2 2 1 100 100 25 75 75
1382.519 2 5 0 100 100 50 0 751795.661 4 4 2 100 100 50 25 751268.476 2 3 2 100 100 50 50 751487.550 3 4 1 100 100 50 50 75
1227.449 3 2 2 100 75 50 75 75S
a
b
S
1071.e7 COOKE ET AL GASTROENTEROLOGY Vol. 137, No. 3
upplementary Table 2. (Continued)
Group m/z [M�Na]
OS composition
Frequency (%)
Pre Post
Hex HexNAc Fucose �4 wk �2 wk 2 wk 8 wk 24 wk
1065.396 2 2 2 75 100 50 75 75976.360 2 3 0 100 100 75 25 75
1138.413 3 3 0 100 100 75 25 751633.608 3 4 2 100 100 75 25 75
814.307 1 3 0 100 100 75 50 751122.418 2 3 1 100 100 75 75 751179.439 2 4 0 100 100 75 75 751430.528 3 3 2 100 100 75 75 751284.471 3 3 1 100 100 100 75 75
Oligosaccharides detected in �25% of monkeys at 2 and 8 weeks postinfection.
Oligosaccharides detected in �50% of monkeys at 2 and 8 weeks postinfection.upplementary Table 3. Acidic Oligosaccharide Mass, Composition, and Frequency of Detection in Antral Gastric BiopsySamples From 4 Monkeys Pre- and Postinfection With H pylori
m/z [M�Na]
OS composition
Frequency (%)
Pre Post
Hex HexNAc Fucose NeuAc NeuGc Na �4 wk �2 wk 2 wk 8 wk 24 wk
1207.3839 4 1 0 1 0 2 0 50 0 0 751264.4054 3 2 0 0 1 2 50 75 0 0 1001410.4633 4 2 0 1 0 2 50 50 0 0 751410.4633 3 2 1 0 1 2 50 50 0 0 751426.4582 4 2 0 0 1 2 75 100 0 0 751572.5161 5 2 0 1 0 2 75 75 0 0 501572.5161 4 2 1 0 1 2 75 75 0 0 501825.6589 1 5 2 1 0 2 75 75 0 0 75
1825.6589 0 5 3 0 1 2 75 75 0 0 75S S
aSorted by Hp� frequency (%).
September 2009 MUCIN OLIGOSACCHARIDES AND H PYLORI INFECTION 1071.e8
upplementary Table 4. Oligosaccharide Mass,Composition, and Frequency ofDetection in Antral Gastric Mucosaof Humans With (Hp�) or Without(Hp�) H pylori Infection
m/z [M�Na]
OS composition Frequency (%)a
Hex HexNAc Fucose Hp� Hp�
611.228 1 2 0 100 100757.285 1 2 1 100 100773.280 2 2 0 100 100814.307 1 3 0 100 100919.338 2 2 1 100 100976.360 2 3 0 100 100
1065.396 2 2 2 100 1001122.418 2 3 1 100 1001138.413 3 3 0 100 1001179.439 2 4 0 100 1001268.476 2 3 2 100 1001284.470 3 3 1 100 1001325.497 2 4 1 100 1001341.492 3 4 0 100 1001430.528 3 3 2 100 1001471.555 2 4 2 100 1001487.550 3 4 1 100 1001544.571 3 5 0 100 1001576.586 3 3 3 100 1001633.608 3 4 2 100 1001690.629 3 5 1 100 1001722.644 3 3 4 100 1001779.666 3 4 3 100 1001795.661 4 4 2 100 1001836.687 3 5 2 100 1001852.682 4 5 1 100 1001941.718 4 4 3 100 1001998.740 4 5 2 100 1002055.761 4 6 1 100 1002087.776 4 4 4 100 1002144.798 4 5 3 100 1002201.819 4 6 2 100 1001649.603 4 4 1 100 882290.856 4 5 4 100 882233.834 4 4 5 100 752306.851 5 5 3 100 752452.909 5 5 4 100 752598.966 5 5 5 100 631747.651 3 6 0 89 1001373.507 3 2 3 89 881211.454 2 2 3 89 752160.793 5 5 2 89 502347.877 4 6 3 89 501982.745 3 5 3 78 881414.533 2 3 3 78 631925.724 3 4 4 78 502509.930 5 6 3 78 502014.735 5 5 1 67 502655.988 5 6 4 67 381738.639 4 3 3 67 252112.783 4 7 0 67 131617.613 2 4 3 67 881227.449 3 2 2 67 1001382.519 2 5 0 56 01081.391 3 2 1 56 88
960.365 1 3 1 56 100
1706.624 4 5 0 44 38upplementary Table 4. (Continued)
m/z [M�Na]
OS composition Frequency (%)a
Hex HexNAc Fucose Hp� Hp�
1884.697 4 3 4 44 252745.024 5 5 6 33 251893.709 3 6 1 33 01017.386 1 4 0 22 381900.692 5 3 3 22 381446.523 4 3 1 22 252363.872 5 6 2 22 252103.771 5 4 3 22 132249.829 5 4 4 22 132964.099 6 6 5 22 132713.009 5 7 3 22 02818.041 6 6 4 22 01592.581 4 3 2 11 50935.333 3 2 0 11 38
1503.545 4 4 0 11 252046.750 5 3 4 11 251300.465 4 3 0 11 131754.634 5 3 2 11 131811.655 5 4 1 11 131868.702 3 3 5 11 131957.713 5 4 2 11 132630.956 7 5 3 11 132379.892 4 4 6 11 02566.952 5 7 2 11 02671.983 6 6 3 11 01763.671 2 4 4 0 381366.524 1 5 1 0 251528.576 2 5 1 0 251535.560 4 2 3 0 251731.656 2 6 1 0 251097.386 4 2 0 0 131243.444 4 2 1 0 131389.502 4 2 2 0 131512.582 1 5 2 0 131608.576 5 3 1 0 132265.824 6 4 3 0 132322.846 6 5 2 0 132395.887 5 4 5 0 132427.877 7 4 3 0 132468.903 6 5 3 0 132557.940 6 4 5 0 132573.935 7 4 4 0 132614.961 6 5 4 0 13